Format

P09. Democratizing Artificial Intelligence: Putting People at the Centre

SCIENCE AND TECHNOLOGY POLICY
PANEL CHAIR(S)
Sushant Kumar
Jindal School of Government and Public Policy
Main chair
Joyeeta Dey
Manipal University
Second chair
CATEGORISATION
POLICY TOPIC
SCIENCE AND TECHNOLOGY POLICY
SECTOR
—
KEYWORDS
—
GENERAL OBJECTIVES, RESEARCH QUESTIONS AND SCIENTIFIC RELEVANCE

Regulation has historically played catch-up with emerging technologies. Awareness is often raised post-facto and in a manner of dealing with risks and benefits emanating from the application of the technology to society at large (Jasanoff 1997, 2016; Law & McCall 2024). The promise of technology alleviating developmental problems often ends up exacerbating them, such as in the case of the Aadhar project in India – envisaged to uniquely identify beneficiaries of transfer services, it has instead resulted in exclusionary effects (Khera 2017). This case in point proves, technology widens the chasms of social, political, and economic inequities, especially for the marginalized, in the name of whom these technological interventions are justified in the first place (Benjamin 2019; Barocas et al., 2017; Moran-Thomas 2020).

The story of development and deployment of Artificial Intelligence (AI) technologies in various walks of life has arguably not been much different. As a technology largely accrued to the future and increasingly permeating the present, AI is often coupled with spectacular promises of both saving mankind as much as posing the risk of rendering it extinct (Future of Life Institute 2023, Stilgoe 2020). Hence, questions of understanding, regulating, and harnessing the technology becomes a critical policy issue.

 

This panel will dive into various policy concerns surrounding AI with an aim to put human developmental objectives at the center of the conversation.

Scientific relevance

The current swathe of AI technologies has proven to be discriminating on grounds of race, class, gender, and even zip codes, a phenomenon popularly known as algorithmic bias (Akter et al., 2021, Stilgoe 2020, Future of Life Institute 2023). They have been proven to further systemic biases encoded in the historical data they’re trained on (Nazer et al., 2023, Gaonkar et al., 2020). Critically analyzing the design, development and deployment of AI technologies across the full stack of this technology, upstream and downstream, and opening this discussion for a wider discussion can help enrich the design of these technologies by making them more inclusive, equitable and accessible and thus enriching democratic values.

Research questions

1.     How are jurisdictions over the world dealing with evolving AI technology?  

2.     How can regulatory frameworks be designed to proactively manage the societal risks and benefits of AI-based technologies to ensure equitable outcomes?

3.     How can historical cases of technological regulation be deployed to inform the development of effective policies for the governance of AI technologies?

4.     How are narratives of hype around the technology influencing policymaking?

5.     How are Governments and the private sector negotiating authority and legitimacy over regulatory aspects of AI technology?

6.     How does geopolitics influence the governance of AI in India and other jurisdictions?

CALL FOR PAPERS

The current swathe of AI technologies has proven to be discriminating on grounds of race, class, gender, and even zip codes, a phenomenon popularly known as algorithmic bias. They have been proven to further systemic biases encoded in the historical data they’re trained on. Historically, regulation has played catch-up with emerging technologies. Awareness is often raised post-facto and in a manner of dealing with risks and benefits emanating from the application of the technology to society at large. The promise of technology alleviating developmental problems often ends up exacerbating them, such as in the case of the Aadhar project – envisaged to uniquely identify beneficiaries of transfer services, it has instead resulted in exclusionary effects. This case in point proves, technology widens the chasms of social, political, and economic inequities, especially for the marginalized, in the name of whom these technological interventions are justified in the first place

 

Given this context, critically analyzing the design, development and deployment of AI technologies across the full stack of this technology, upstream and downstream, and opening this discussion for a wider discussion can help enrich the design of these technologies by making them more inclusive, equitable and accessible and thus enriching democratic values.

 

This panel will dive into various policy concerns surrounding AI with an aim to put concerns of justice, equity, and access at the center of the conversation.

 

 We invite contributions from a range of fields, practices, and individuals who speak to to one or more of the following research questions:

 Research questions

1.     How are jurisdictions over the world dealing with evolving AI technology?  

2.     How can regulatory frameworks be designed to proactively manage the societal risks and benefits of AI-based technologies to ensure equitable outcomes?

3.     How can historical cases of technological regulation be deployed to inform the development of effective policies for the governance of AI technologies?

4.     How are narratives of hype around the technology influencing policymaking?

5.     How are Governments and the private sector negotiating authority and legitimacy over regulatory aspects of AI technology?

6.     How does geopolitics influence the governance of AI in India and other jurisdictions?

 

Types of papers expected, including but not limited to:

 

1.        Anticipatory governance of AI technologies

2.        Implications for caste, class, gender, linguistic, technological and regional divides

3.        Use cases from policy sectors including but not limited to health, education, urban governance

4.        Narratives, discourse, framing and communication

5.        Geopolitics and AI nationalisms

 

Please submit a paper title and abstract (200-300 words) by the 31st of January, 2025 to the ICPP7 conference site. We look forward to receiving your submissions and engaging in lively discussions about regulating and democratizing Aritificial Intelligence technologies.

 

ACCEPTED
Format
01 - Effective Implementation of Healthcare AI: Lessons from Six Frontier Technology Pilots in Developing Contexts
Christopher Moore (University College London, University of London)
CO-AUTHORS
—
ABSTRACT

As the world of innovative technologies grows to address disparities in health systems, AI-based health applications are being increasingly employed in new contexts. Technologies such as AI-based health tools and chatbots may support gaps in resources and expertise with the aim to increase access to reliable and high-quality healthcare services. Their use is continuing to expand—particularly due to the COVID-19 pandemic—from global minority to now global majority contexts, such as in LMICs.


With this growth comes a broad range of challenges and considerations for effective use, from access to affordable technologies to trustworthiness in new, non-traditional systems. Given the lack of evaluative studies to support decision-making surrounding AI and Machine-Learning use in low-resourced settings, this research explores six pilots that utilized AI for health in developing contexts to identify any differences or commonalities in their design and delivery and what lessons could be gleaned for wider adoption and implementation of AI for health tools in the developing world.

 

This research looked at six pilots delivered by the UK-based Frontier Technologies Hub, through interviews with key actors within the pilots, including pioneers, technologists and innovation coaches. All included pilot programs used AI in some capacity to accomplish health goals in developing contexts, highlighting the differences or commonalities in the program designs and delivery and what lessons could be gleaned for wider adoption and implementation of AI for health tools in the developing world. The paper discusses the results from the interviews with key actors and the actionable insights into how AI tools can be more effectively designed, implemented, and scaled for sustainable health outcomes in developing regions.


We highlight here that the research in this topic is at a systematic level or applies Western ideas and research to developing contexts and questions. There is a clear gap in the literature for analysis on real-world attempts to utilise and integrate AI technologies in healthcare systems within developing contexts. This research supports future innovators to develop their own plans or pilots for similar schemes as well as begin to build a research base for practical implementation of AI in these contexts that is currently not present. By synthesizing lessons learned, this paper aims to empower global health stakeholders—policymakers, NGOs, and tech innovators—with strategies to bridge the gap between technological promise and practical impact.

ACCEPTED
Format
02 - Making national AI sovereignty against global data capitalism? The case of India
Ritam Sengupta (O.P. Jindal Global University, India)
CO-AUTHORS
—
ABSTRACT

At least since the launch of GPT 4, that gave widescale access to Artificial Intelligence applications to the lay public, the mediation of everyday life by AI has become a palpable reality in the Indian subcontinent. My paper will explore how the governance of AI in India is particularly geared towards the question: how can AI be nationalised/localised/channelled into the crafting of techno-economic sovereignty? In pursuing this problematic, I will look at the conception of such techno-economic sovereignty in the formulation of state policy and in statements by state functionaries who bear a long-standing anxiety about the global nature of data capitalism. I will try to trace how the reproduction of such anxieties around issues of AI enables real and virtual enclosures – of the production of data and its (cultural) uses - within which data-based applications can be assigned a national, autochthonous career. This will be a comparative take with an eye towards legislative developments in the UK, Europe and USA about AI dealing with similar issues. On a second register, I will study how global and local intersections matter to a more people-centric understanding of the use of AI in contemporary India, with the caveat that most of the AI applications in this period are run through already silo-ed worlds of multinational corporations. Finally I will try to understand what are the implications of the imagination of (nationally) sovereign AI and engagement with global data capitalism for a possibly more democratic vision of AI use in contemporary India.

ACCEPTED
Format
03 - Understanding the democratisation of AI governance as a complex adaptive process
Naresh Singh (Jindal School of Government and Public Policy)
CO-AUTHORS
—
ABSTRACT

This paper argues that the governance and regulation of AI must be conceptualized as a complex adaptive systems (CAS) process, if its rapid technological evolution and equally rapid spread in all strata of society, can be sensibly guided to the greater good of the greatest number of people. CAS theory studies systems made up of interconnected dynamic components that adapt and evolve in response to changes in the environment. This is exactly the kind of situation that is playing out with the coevolution of AI technological growth and its interaction with society. As such it is proposed that this kind of thinking can help create governance arrangements that are resilient, participatory and capable of evolving to meet human needs as these also continue to evolve.

 

After a review of existing AI regulations in selected countries, the paper will show how CAS principles such as coevolution, adaptation, self-organization, emergence, non-linearity, and decision making in the face of uncertainty, can help democratize AI governance. To make this operational, processes and mechanisms such as participatory feedback mechanisms, AI monitoring systems, inclusion of diverse agents in the social system including marginalized voices, crowd sourcing platforms to foster emergent collaboration and use of agile governance structures to deal with non-linear unanticipated impacts will be discussed.

 

The paper then considers concrete recommendations for policy makers. These include the creation of adaptive and iterative policy processes which evolve in response to AI development and feedback from the public. To facilitate this the use of policy thinktanks and regulatory sandboxes can be used to test run regulations before broader deployment. Learning must become the key management strategy and be supported by open transparent and accountable feedback systems which create trust while protecting privacy.

 

India is the first country to have attempted to articulate this approach in a modest manner so far and the paper will conclude with a brief review of the approach as an illustration.

 

ACCEPTED
Format
04 - Entangled Dependencies: AI, Global Capitalism, and the Neoliberal Order in Developing Economies
Chandraketan Sahu (Jindal Global Business School)
CO-AUTHORS
—
ABSTRACT

The proliferation of artificial intelligence (AI) technologies in developing economies is deeply embedded within the global technology ecosystem, through their systemic dependency on capital, technology, and infrastructure primarily controlled by the global north. Through the ongoing narratives of free market globalization, open innovation, and digital development, the elements of this technology ecosystem shape the institutional infrastructure of these economies. This dependency is further amplified by localized adoption of emerging business models, practices, innovations, along with prescriptive modes of operations andorganizing. Simultaneously normalized is the structure of this dependency, where the role of developing economies in global south is to provide an open expanded market for finished technology products while also being a source of labor and entrepreneurs who can mobilize the borrowed capital and technology to innovate new products for the local markets. Upon careful examination, the roots of this dependency can be traced to historically entrenched power dynamics within global capitalism, which has evolved into the current operational structures of the capitalist neoliberal internet economy. Building on foundational critiques of global capitalism (Wallerstein, 1974; Harvey, 2005) and decolonial theory (Quijano, 2000; Escobar, 1995), this research interrogates the historical and institutional underpinnings of these dependencies. 
 
Previous studies have explored the dynamics of technology diffusion and digital development, yet limited attention has been given to the interplay of geopolitics, neoliberal capitalism, and sociotechnical imaginaries in shaping the trajectories of AI adoption and governance in the global south (Jasanoff & Kim, 2009, Kuchler & Stigson, 2024). This research addresses this concern by critically analyzing the narratives, practices, and structures that sustain these dependencies. The global AI landscape is structured around dominant narratives of free-market globalization, open innovation, and digital development that ostensibly promise inclusivity and equality but often mask the perpetuation of extractive relationships. This study reveals how these narratives co-opt local adoption of emerging business models, innovations, and practices in developing economies while prescribing modes of organizing that align with the profit-maximizing imperatives of global capitalism (Couldry & Mejias, 2019). By situating AI development within the historical context of colonial enterprise, this research underscores how developing economies in global South, are positioned as open markets for finished technological products while simultaneously serving as sources of labor and entrepreneurial talent for mobilizing borrowed technologies for local contexts. 
 
This research employs critical discourse analysis of texts, media reports, and corporate narratives surrounding TikTok and Huawei, two Chinese corporations that have challenged the dominance of Western firms. Drawing on various media sources (NYT, Economist, and Financial Times), as well as scholarly critiques of digital capitalism (Schiller, 1999), the study examines the systematic vilification of these firms. This discourse, it is argued, is not solely an economic maneuver to protect Western market dominance but also a geopolitical strategy to neutralize potential shifts in the global centers of accumulation, extraction, and profitability. The case of India, where TikTok was banned under the pretext of national security amid ongoing border tensions, exemplifies how such measures are often employed to entrench the dominance of American platforms like YouTube, Instagram, and Twitter, which now monopolize India’s social information and communication market (Bhattacharya, 2021). Even when firms from alternative locations demonstrate capacity to compete on innovation, quality, and accessibility, e.g. (TikTok and Huawei), they face intervention through sovereign state concerns, such as national security, that align with broader Western imperial interests (Singh & Chander, 2022). Furthermore, the debates on data colonialism, global governance, and AI ethics are shown to reinforce, rather than disrupt, these dynamics. Concepts of inclusivity, diversity, and equality are often co-opted to create a veneer of legitimacy while preserving the core-periphery structures of global capitalism.
 
 
The results of this research reveal that the global technology ecosystem is a deeply hierarchical structure where access to markets, resources, and legitimacy is mediated by Western-centric norms and geopolitical considerations. 

ACCEPTED
Format
05 - Generating Democracy: Mapping the horizons of AI and core electoral processes
Akansha Yadav (O.P. Jindal Global University)
CO-AUTHORS
—
ABSTRACT

In this paper we have dwelled into exploring out the interaction of AI with the electoral process, considering this is the key component of democracy, it becomes pertinent to understand the exploitative nature of AI which is being misused to undermine the democratic principles. We have chosen the jurisdictions of United States of America, European Union and India for this paper. It Is well documented that there has been significant enthusiasm across the globe in respect of using AI for all forms of social activity. However, the electoral process—the time, place, and manner of elections within democratic nations—is one of few sectors in which there has been limited penetration of AI. Electoral management bodies in many countries have recently started exploring and deliberating over the use of AI in the electoral process.  Artificial Intelligence (AI) has emerged as a transformative force in political campaigns and elections, significantly influencing freedom of expression through algorithmic content curation, micro-targeting, and automated decision-making systems. While AI has the potential to enhance democratic participation, its unchecked deployment poses serious risks to political discourse, media pluralism, and electoral integrity. The absence of transparency and accountability mechanisms in AI-driven content moderation and promotion has led to the amplification of disinformation, suppression of marginalized voices, and manipulation of voter perceptions.

The self-regulation model adopted by digital platforms has proven inadequate in addressing these challenges, necessitating a robust regulatory framework to ensure AI operates in alignment with human rights principles. The European Commission’s proposed Digital Services Act (DSA) and Digital Markets Act (DMA) mark an important step toward regulating online platforms, but additional measures are required to ensure transparency, accountability, and fairness in AI-driven political communication. A comprehensive legal framework should mandate independent audits of AI systems, public access to data for research purposes, and standardized content moderation guidelines developed in collaboration with civil society organizations.

The democratization of AI is imperative to prevent monopolization by a few tech conglomerates and ensure equitable access to AI tools for political actors, journalists, and civil society. Open-source AI solutions, interoperability standards for content moderation, and public oversight mechanisms should be prioritized to counterbalance the power of private entities in shaping political narratives. Additionally, the regulation of micro-targeting and digital political advertising must be strengthened to prevent manipulation and ensure transparency in electoral processes. A global regulatory approach, rooted in international human rights standards, is needed to address the cross-border nature of AI-driven political influence. Governments, policymakers, and tech companies must collaborate to create a governance model that upholds democratic values, safeguards freedom of expression, and promotes an open, fair, and inclusive digital space. Without proactive intervention, AI risks becoming a tool for digital authoritarianism rather than a force for democratic empowerment.

In this paper, we consider five avenues within the core electoral process which have potential for AI usage, and map the challenges involved in using AI within them. These five avenues are: voter list maintenance, determining polling booth locations, polling booth protection processes, voter authentication, and video monitoring of elections. Within each avenue, we lay down the context, illustrate current or potential usage of AI, and discuss extant or potential ramifications of AI usage, as well as potential directions for mitigating risks when considering AI usage. We believe that the scant current usage of AI within electoral processes provides a very rare opportunity to deliberate on the risks and mitigation possibilities prior to actual and widespread AI deployment. This paper is an attempt to map the horizons of risks and opportunities in using AI within electoral processes and to help shape the debate around the topic.

 

Keywords: Artificial Intelligence, Elections, Corrupt Practices, Democracy, Electoral Process

ACCEPTED
Format
06 - AI as a More Democratic Pathway – Social Media, Communicative Rationality, and Public Policy
Gaajarla Sheshugoud (Osmania University, Hyderabad )
CO-AUTHORS
—
ABSTRACT

Artificial Intelligence (AI) is transforming public policy formulation by making governance more participatory, transparent, and responsive. In the digital age, social media platforms serve as an open forum for public discourse, where citizens engage in policy discussions, critique government actions, and advocate for change.Traditional policymaking has often been a top-down process, where decisions are driven by bureaucratic and political considerations, often ignoring the voices of marginalized communities. AI, through natural language processing (NLP) and sentiment analysis, enables governments to analyze large-scale social media interactions, extract key themes, and assess public sentiment in real time. This data-driven approach ensures that policies are aligned with public concerns rather than elite preferences, making AI a critical tool for democratising governance.

 

A fundamental component of AI-driven policy formulation is text as “communicative rationality”, where public discussion and reasoning contribute to governance decisions. Drawing from Jürgen Habermas' theory of communicative action, this approach emphasises that rational discourse, rather than political coercion, should shape public policies. AI facilitates this process by analysing millions of comments, petitions, and discussions across digital platforms, allowing policymakers to identify dominant narratives, public grievances, and social priorities. By translating digital discourse into actionable insights, AI ensures that governance becomes more deliberative and inclusive. Governments can now use AI-powered discourse mapping to identify policy gaps, understand the effectiveness of existing measures, and predict public reactions before policy implementation.

 

The objective of this study is to explore how AI-driven analysis of social media discourse enhances participatory governance and contributes to a more democratic public policy approach. To achieve this, I will examine how AI processes public discussions using NLP-based sentiment analysis and opinion mining, providing a quantitative and qualitative understanding of policy-related conversations. Additionally, I will analyze real-world case studies where AI-driven public sentiment analysis influenced governance decisions in areas like development and displacement, political representation and recruitment , and social welfare programs. Furthermore, I will investigate how AI-powered feedback loops enable continuous policy adaptation, ensuring that public policies remain dynamic and reflective of evolving citizen needs.

 

While AI presents vast opportunities for democratic governance, challenges such as algorithmic bias, misinformation, and ethical concerns must be addressed to prevent AI-driven policymaking from reinforcing existing inequalities. This paper will also explore frameworks for responsible AI governance, ensuring that AI remains an enabler of democracy rather than a tool for surveillance or manipulation. As public discourse continues to shift to digital platforms, AI-powered policymaking will serve as a bridge between citizens and governments, fostering a governance model that is more inclusive, responsive, and democratic.

 

ACCEPTED
Format
07 - Towards Responsible AI in India: A Strategic Framework for Robust Governance and Impact Assessment of AI Systems
Pranjal Khare (O.P. Jindal Global University)
CO-AUTHORS
Paridhi Sharma (OP Jindal Global University, Sonipat )
ABSTRACT

Towards Responsible AI in India: A Strategic Framework for Robust Governance and  Impact Assessment of AI Systems

This paper will delve into the transformative impact of Artificial Intelligence (AI) across various sectors while highlighting the ethical, legal, and social challenges arising from its rapid advancement. AI has demonstrated immense potential in revolutionising industries, enhancing efficiency, and contributing to societal progress. However, these benefits also raise significant concerns about fairness, accountability, transparency, privacy, security, and adherence to human-centric values. Addressing these issues is crucial to ensuring that AI development and deployment align with the principles of Responsible AI.

To facilitate the responsible governance of AI technologies, it is imperative to establish a comprehensive and structured "AI Impact Assessment Mechanism" tailored to the Indian context. This mechanism would serve as a regulatory and evaluative framework to systematically assess AI systems and applications, ensuring they conform to ethical standards, legal mandates, and policy guidelines. By implementing such an assessment mechanism, policymakers and regulatory authorities can gain a thorough understanding of the AI ecosystem in India. This will aid in mitigating potential risks while maximising AI's benefits in a controlled and accountable manner.

In recent years, India has taken significant steps toward AI governance, particularly through the formulation of the National Strategy on AI (2018) and its adherence to Responsible AI principles endorsed by global organizations such as the Organisation for Economic Co-operation and Development (OECD) and the United Nations Educational, Scientific and Cultural Organization (UNESCO). These policy frameworks emphasize the necessity of fairness, inclusivity, and ethical considerations in AI applications. Despite these efforts, the absence of a structured impact assessment mechanism leaves a regulatory gap that must be addressed to ensure AI technologies are deployed in a manner that aligns with national and international best practices.

The proposed "AI Impact Assessment Mechanism" aims to bridge this gap by integrating a multi-dimensional evaluation process that scrutinizes AI applications through legal, ethical, and societal lenses. The mechanism would involve key stakeholders, including government agencies, AI developers, industry experts, and civil society organizations, to create a balanced and inclusive approach toward AI governance. This framework would facilitate risk assessment, regulatory compliance checks, and continuous monitoring of AI systems to ensure they adhere to India's existing legal framework and ethical norms.

Additionally, this assessment model would empower the government with a structured oversight mechanism that fosters trust and accountability in AI deployment. By enforcing compliance with ethical AI principles, it would contribute to reducing bias, preventing misuse, and promoting transparency in AI-driven decision-making processes. The AI Impact Assessment Mechanism would also enhance public confidence in AI technologies by ensuring their implementation prioritizes social well-being and adheres to human rights standards.

 

Keywords: - Artificial Intelligence; Assessment; Co-operation and Development (OECD); United Nations Educational, Scientific and Cultural Organization (UNESCO)

 

ACCEPTED
Format
08 - An Explanatory Assessment of the Strategies to Promote the Participation of Mauritian Women In Artificial Intelligence
Toolsy Luchmun (Open University of Mauritius)
CO-AUTHORS
—
ABSTRACT

This research investigates strategies to promote the participation of Mauritian women in Artificial Intelligence (AI) within the context of the country’s emerging AI landscape. The primary research question is: What is Mauritius doing to promote the participation of Mauritian women in AI? Despite the growing momentum of AI development in Mauritius, there is limited research on the participation of women in the sector.

 

The study breaks down the main question into four sub-questions: (1) What is the status of AI in Mauritius (with a special focus on gender)? It assesses the current AI landscape in Mauritius, while exploring gender disparities within the sector. It probes whether women face marginalisation or if AI-related incentives are offered. It evaluates whether Mauritian women have equal opportunities as their male counterparts to rise as leaders in AI development and implementation.

 

(2) What challenges do Mauritian women face in pursuing careers in AI? It evaluates the challenges faced by secondary schools in guiding and supporting girls towards AI career paths and assesses the barriers female students encounter at the tertiary level. It also evaluates how well the work environment supports women in AI-related fields.

 

(3) What strategies currently exist to promote their participation? Investigating existing policies and initiatives will help evaluate current efforts and identify areas for improvement. (4) What additional measures can be adopted to enhance their engagement? It examines measures to be taken or enforced at the secondary school level to guide girls toward AI careers, at the tertiary level to encourage women to pursue AI, workplace initiatives to support women in AI, and collaborative policies between the government and private sector to promote women’s participation in AI.

 

A concurrent mixed-methods approach is employed, combining quantitative data from 80 professionals in the ICT and AI sectors with qualitative insights from secondary research. Statistical analysis using SPSS, including descriptive statistics, t-tests for gender differences as well as correlation analyses, would be conducted. The qualitative component involves thematic analysis of existing AI-related policies and academic literature.

 

Gender is crucial to democratization. This research aligns with the panel theme Democratising Artificial Intelligence: Putting People at the Centre and addresses the panel’s research question: How can regulatory frameworks be designed to proactively manage the societal risks and benefits of AI-based technologies to ensure equitable outcomes? By focusing on Mauritius, the first African country to adopt a national AI policy, this study provides insights into how gender-inclusive AI frameworks can be developed in the African context. Notably, despite Mauritius being a leader in Africa's IT sector, the National AI Strategy and the Digital Mauritius 2030 Strategic Plan, which guide AI development, both lack specific measures to promote women’s participation in AI.

 

Additionally, the World Economic Forum (2023) highlights that women are mostly in the entry-level roles in STEM fields, which is why this panel has been chosen – to emphasise the need for women to be at the center of AI development in democratic governance.

 

It should be noted that a substantial amount of research and data have already been collected for this study and just needs to be finalised into a paper.

ACCEPTED
Format
09 - AI for All? Bridging Gender Gaps in the Age of Democratized AI
Sunidhi Setia (OP Jindal Global University, Jindal Global Law School)
CO-AUTHORS
Aditya Agrawal (OP Jindal Global University)
ABSTRACT

With Large Language Models or LLMs and Generative AI, such as ChatGPT and Deepseek, taking over the world, the question of internet/AI governance becomes vital. Since AI models are trained on existing literature, they are implicitly and unsuspectingly trained on biased material creating critical issues of gender imbalance. Therefore, in order to have an inclusive and equitable technological system, it becomes important to accept that there exist an implicit or unintentional bias. To understand the tech gender bias, the translation from one language to another of gender neutral terms is noteworthy. The results will show that the Language processing models are associating the certain professions, behaviours or roles with specific genders. The example reinforces the programmers’ biases or skewed patterns inflicted on machines while feeding training data. The huge diversity in the answers to series of sensitive geo-political questions posted to both famous ChatGPT and newly launched Deepseek by the news platform inshorts under insights by inshorts on January 31, 2025 is yet another recent depiction of transfer of human biases to machine through data transfer. Shifting from the abstract theoretical implementation of AI, the paper is advocating inclusive, equitable and diverse approach for AI creation and implementation. With feminist analysis of current AI policies the research seeks to redefine the intersection of technology, gender and society. While highlighting the advent of school of Feminist Artificial Intelligence the paper critically explore the historical exclusion of women’s knowledge and perspective in scientific discourse. Promoting Alison Adam’s critique of AI, the authors doesn’t promote FAI as an alternate but as a means to well-informed and inclusive AI. Signifying the trajectory and progression for an inclusive AI, the research focuses on importance of Models, design and data collection that can be very well envisaged in progression of conceptual framework to tangible applications.

 

The research centrally argue for development of AI technologies that stands as a liberating force representing feminist theory. With AI slowly seeping into all the major sectors around the world from healthcare to legal aid to pharma to food industry, it’s imperative to calculate the cost of tech discrimination to society as whole. The authors realising the risk associated with such biases provides a three pronged feminist approach to the AI development and implementation. This research through a performative model of awareness, inclusion and regulation stresses for a more democratic policy making to bridge the gendered gap. Being cognizant of the importance of broader initiatives and collective solidarity to mitigate such discrimination the research suggests mandatory tech auditing and sensitization to have an inclusive AI transformation.

ACCEPTED
Format
10 - The Need for AI Regulation and Ethical Oversight
Akansha Yadav (O.P. Jindal Global University)
CO-AUTHORS
Aditya Agrawal (OP Jindal Global University)
ABSTRACT

Artificial Intelligence (AI) has rapidly evolved from a futuristic concept into an indispensable tool in various industries, reshaping economies, governance, and everyday life. As its influence expands, prominent figures in the technology sector, including Sundar Pichai, Elon Musk, and Brad Smith, have called for proactive regulation to mitigate its risks. AI-powered applications such as deep fakes, facial recognition, and autonomous systems pose significant threats to public safety, privacy, and democratic institutions. The debate over AI regulation is multifaceted, encompassing ethical, legal, and geopolitical concerns.

A central question in the AI discourse is whether AI should be regulated at all, and if so, who should oversee its governance? While proponents of regulation argue that unchecked AI development could lead to existential threats—including the misuse of autonomous weapons and AI-driven disinformation campaigns—opponents caution against premature regulatory intervention that could stifle innovation. Azamat Abdoullaev contends that AI, as a fundamental technology, should remain free from early-stage regulations, as improper oversight may hinder its growth. However, figures like Stephen Hawking and Bill Gates warn that an unregulated AI landscape could lead to disastrous consequences, emphasizing the urgent need for ethical and legal safeguards. The regulatory dilemma presents two contrasting approaches: state-imposed regulation versus self-regulation by industry players. Government-led regulations offer a structured framework to ensure accountability and fairness, but excessive control may impede technological progress and lead to uneven global competitiveness. In contrast, self-regulation relies on corporate responsibility, which, while fostering innovation, risks creating an environment where ethical concerns are overlooked in the pursuit of market dominance. A balanced approach, incorporating mandatory ethical standards and sector-specific oversight, is crucial to navigating AI governance effectively.

India's stance on AI regulation, as outlined in NITI Aayog's ‘National Strategy for Artificial Intelligence’ and the 2020 Working Document on Responsible AI, underscores the need for a flexible, risk-based regulatory framework. The proposed oversight body, comprising experts from various disciplines, aims to develop sector-specific guidelines, ensuring AI deployment aligns with national interests and ethical considerations. However, the absence of a comprehensive AI-specific legal framework in India, as noted by legal scholars, highlights gaps in policy execution and enforcement. Internationally, efforts such as the European Union's Digital Services Act (DSA) and Digital Markets Act (DMA)exemplify legislative attempts to regulate AI-driven platforms, ensuring accountability, transparency, and user rights protection. However, the lack of global consensus remains a significant hurdle, as divergent regulatory frameworks across countries create inconsistencies in AI governance.

This research paper will examine the need for AI regulation, ethical dilemmas, privacy concerns, and security threats posed by unregulated AI. Secondly we have discussed the key stakeholder perspectives, analysis of views from technology leaders, policymakers, legal scholars, and AI ethicists. We have also discussed the global regulatory approaches for which we have included case studies from the EU, the US, and India, exploring existing and proposed AI governance models. Authors have also tried to address the challenges in AI regulation and have attempted to provide the future roadmap for AI governance for a globally coordinated regulatory framework, emphasizing transparency, accountability, and ethical AI development.

 

Keywords: Artificial Intelligence, Humanised AI, Democracy, Public Policy, Responsible AI.

ACCEPTED
Format
11 - The Protagonists of AI Narratives: An NLP Analysis of Stakeholder Perspectives in the Indian AI Landscape
Purnima Rohilla (Indian Institute of Technology, Bombay)
CO-AUTHORS
Anupam Guha (Indian Institute of Technology, Bombay)
ABSTRACT

As Artificial Intelligence (AI) rapidly transforms various sectors, from finance, healthcare, and government functions to transportation, manufacturing, and education, governments worldwide are developing policies for its adoption. These policies are crucial in determining how society adopts AI technology (Guenduez & Mettler, 2023). Other stakeholders, including private companies, media, research institutions, and think tanks, also significantly shape the AI policy landscape (Deshpande & Sharp, 2022). The stakeholders often influence policy processes through narratives and agenda-setting. 

Hayes & Fitzpatrick (2024) argue that narratives inform and structure relations between people and technology, and dominant groups are better positioned to control these narratives. Numerous stakeholders, including tech companies and management consultancies, regularly employ sophisticated public relations campaigns to position AI as an indispensable and trustable part of the future using whitepapers, surveys, and reports (Bourne, 2019).

The Narrative Policy Framework (NPF) is a theoretical approach which analyses how policy narratives shape public opinion and policy outcomes. NPF is a useful method for understanding stories that support and reinforce the norms and rules of institutions. The Framework asserts that "policy narratives have some combination of a setting, characters (heroes, victims, and villains), plots, and a moral of the story (policy solution)." These are the four "key elements" of narratives (Jones et al., 2014). In the context of AI policy, NPF can help identify and analyse the narratives constructed by the stakeholders to understand their impact on AI governance. This technique lends itself to analysis by NLP tools, historically used for narrative analysis. In particular, temporal topic modelling like Bayesian hidden topic Markov model (Gruber et al., 2007) and Relationship Modeling Networks (Iyyer et al., 2016) have been used to analyse character relationships in stories. 

Our work uses these NLP methods combined with qualitative discourse analysis to analyse government documents, policy papers, statements, and reports from private companies, newspaper articles, reports, and research publications. Furthermore, this research clarifies the scope, significance, and consequences of using machine learning in social research. Expanding the scope of research in these directions can provide a more nuanced understanding of AI governance and the role of stakeholders in shaping public opinion and policy.

ACCEPTED
Format
12 - What is AI democratization?
Fernando Filgueiras (Federal University of Goiás)
CO-AUTHORS
—
ABSTRACT

The pervasiveness and ubiquity of AI has drawn attention to the potential dangers and threats that algorithmic systems pose in different parts of society. Artificial intelligence disruptively transforms decision-making and the performance of various tasks in society. A rigorous body of literature draws attention to potential threats such as injustice, discrimination, exclusion, and various organizational risks. In response to these AI threats, the literature places democracy as essential. An emerging body of literature has constructed claims for AI democratization, without precisely defining what the meaning of democracy is with which we seek democratization. The common thread in this emerging struggle is that democratizing AI means expanding society’s access to tools, data sets, algorithms, and computing power. In this article, we argue that the meaning of AI democratization goes beyond this issue. The problem with AI democratization is how we institutionalize the decision-making process involving complex interactions between humans and artificial agents. In this sense, AI democratization must be thought of within emerging algorithmic regimes. The objective of this article is to analyze the concept of AI democratization and to consider possible frameworks for the insertion of AI in the dynamics of democratic political regimes. 

ACCEPTED
Format
13 - Biometric verification for education administration
Joyeeta Dey (Manipal University)
CO-AUTHORS
Sushant Kumar (Jindal School of Government and Public Policy)
ABSTRACT

Across the world  biometric systems are increasingly being adopted by States like India, Nigeria, United States and UAE,for more ‘secure, accurate, efficient’ governance. In the domain of education administration as well, biometric systems have several use cases ranging from monitoring teacher attendance to ascertaining ‘real’ numbers in terms of student enrolment by weeding out false or duplicate identity claims. In this paper I will go over prominent use cases across the world, proceeding to then theorise both the discourse they are embedded in as well as their materiality using concepts from Science and Technology Studies.

Science and Technology studies serves as an especially fruitful conceptual toolbox to navigate understanding of these processes. I will tracing how the idea of ‘truth’ is attached to the informationalised body and then this bio-information isused as a state sanctioned sorting deviceto distribute benefits/rewards to the ‘deserving’. Infrastructure surrounding the public delivery of services is built on the affordance of certainty and the claim of immunity to proxies this technology provides.

It is also useful to view it as an obligatory passing point - both for the state and the citizen, alongside tech companies, civil society and other intermediaries who form an assemblage to produce and stabilise this infrastructure of truth making.In addition, each actor in this assemblage of the socio-technical artifact has diverse meanings and narratives about it, some overlapping some discrete, some even contradictory. Such socio-technical artifacts embody state making projects, holding a special significance for developing and postcolonial societies. In the latter contexts there is a tendency to slide into uncritical triumphalism with regard to participation in modernity through technological adoption. Therefore, in these paper I will also critically read these narratives to evaluate whose perspectives are adopted and whose get marginalised, querying for questions of state surveillance, privacy and exclusion.

CO-AUTHORS
—
ABSTRACT

As the world of innovative technologies grows to address disparities in health systems, AI-based health applications are being increasingly employed in new contexts. Technologies such as AI-based health tools and chatbots may support gaps in resources and expertise with the aim to increase access to reliable and high-quality healthcare services. Their use is continuing to expand—particularly due to the COVID-19 pandemic—from global minority to now global majority contexts, such as in LMICs.


With this growth comes a broad range of challenges and considerations for effective use, from access to affordable technologies to trustworthiness in new, non-traditional systems. Given the lack of evaluative studies to support decision-making surrounding AI and Machine-Learning use in low-resourced settings, this research explores six pilots that utilized AI for health in developing contexts to identify any differences or commonalities in their design and delivery and what lessons could be gleaned for wider adoption and implementation of AI for health tools in the developing world.

 

This research looked at six pilots delivered by the UK-based Frontier Technologies Hub, through interviews with key actors within the pilots, including pioneers, technologists and innovation coaches. All included pilot programs used AI in some capacity to accomplish health goals in developing contexts, highlighting the differences or commonalities in the program designs and delivery and what lessons could be gleaned for wider adoption and implementation of AI for health tools in the developing world. The paper discusses the results from the interviews with key actors and the actionable insights into how AI tools can be more effectively designed, implemented, and scaled for sustainable health outcomes in developing regions.


We highlight here that the research in this topic is at a systematic level or applies Western ideas and research to developing contexts and questions. There is a clear gap in the literature for analysis on real-world attempts to utilise and integrate AI technologies in healthcare systems within developing contexts. This research supports future innovators to develop their own plans or pilots for similar schemes as well as begin to build a research base for practical implementation of AI in these contexts that is currently not present. By synthesizing lessons learned, this paper aims to empower global health stakeholders—policymakers, NGOs, and tech innovators—with strategies to bridge the gap between technological promise and practical impact.

CO-AUTHORS
—
ABSTRACT

At least since the launch of GPT 4, that gave widescale access to Artificial Intelligence applications to the lay public, the mediation of everyday life by AI has become a palpable reality in the Indian subcontinent. My paper will explore how the governance of AI in India is particularly geared towards the question: how can AI be nationalised/localised/channelled into the crafting of techno-economic sovereignty? In pursuing this problematic, I will look at the conception of such techno-economic sovereignty in the formulation of state policy and in statements by state functionaries who bear a long-standing anxiety about the global nature of data capitalism. I will try to trace how the reproduction of such anxieties around issues of AI enables real and virtual enclosures – of the production of data and its (cultural) uses - within which data-based applications can be assigned a national, autochthonous career. This will be a comparative take with an eye towards legislative developments in the UK, Europe and USA about AI dealing with similar issues. On a second register, I will study how global and local intersections matter to a more people-centric understanding of the use of AI in contemporary India, with the caveat that most of the AI applications in this period are run through already silo-ed worlds of multinational corporations. Finally I will try to understand what are the implications of the imagination of (nationally) sovereign AI and engagement with global data capitalism for a possibly more democratic vision of AI use in contemporary India.

CO-AUTHORS
—
ABSTRACT

This paper argues that the governance and regulation of AI must be conceptualized as a complex adaptive systems (CAS) process, if its rapid technological evolution and equally rapid spread in all strata of society, can be sensibly guided to the greater good of the greatest number of people. CAS theory studies systems made up of interconnected dynamic components that adapt and evolve in response to changes in the environment. This is exactly the kind of situation that is playing out with the coevolution of AI technological growth and its interaction with society. As such it is proposed that this kind of thinking can help create governance arrangements that are resilient, participatory and capable of evolving to meet human needs as these also continue to evolve.

 

After a review of existing AI regulations in selected countries, the paper will show how CAS principles such as coevolution, adaptation, self-organization, emergence, non-linearity, and decision making in the face of uncertainty, can help democratize AI governance. To make this operational, processes and mechanisms such as participatory feedback mechanisms, AI monitoring systems, inclusion of diverse agents in the social system including marginalized voices, crowd sourcing platforms to foster emergent collaboration and use of agile governance structures to deal with non-linear unanticipated impacts will be discussed.

 

The paper then considers concrete recommendations for policy makers. These include the creation of adaptive and iterative policy processes which evolve in response to AI development and feedback from the public. To facilitate this the use of policy thinktanks and regulatory sandboxes can be used to test run regulations before broader deployment. Learning must become the key management strategy and be supported by open transparent and accountable feedback systems which create trust while protecting privacy.

 

India is the first country to have attempted to articulate this approach in a modest manner so far and the paper will conclude with a brief review of the approach as an illustration.

 

CO-AUTHORS
—
ABSTRACT

The proliferation of artificial intelligence (AI) technologies in developing economies is deeply embedded within the global technology ecosystem, through their systemic dependency on capital, technology, and infrastructure primarily controlled by the global north. Through the ongoing narratives of free market globalization, open innovation, and digital development, the elements of this technology ecosystem shape the institutional infrastructure of these economies. This dependency is further amplified by localized adoption of emerging business models, practices, innovations, along with prescriptive modes of operations andorganizing. Simultaneously normalized is the structure of this dependency, where the role of developing economies in global south is to provide an open expanded market for finished technology products while also being a source of labor and entrepreneurs who can mobilize the borrowed capital and technology to innovate new products for the local markets. Upon careful examination, the roots of this dependency can be traced to historically entrenched power dynamics within global capitalism, which has evolved into the current operational structures of the capitalist neoliberal internet economy. Building on foundational critiques of global capitalism (Wallerstein, 1974; Harvey, 2005) and decolonial theory (Quijano, 2000; Escobar, 1995), this research interrogates the historical and institutional underpinnings of these dependencies. 
 
Previous studies have explored the dynamics of technology diffusion and digital development, yet limited attention has been given to the interplay of geopolitics, neoliberal capitalism, and sociotechnical imaginaries in shaping the trajectories of AI adoption and governance in the global south (Jasanoff & Kim, 2009, Kuchler & Stigson, 2024). This research addresses this concern by critically analyzing the narratives, practices, and structures that sustain these dependencies. The global AI landscape is structured around dominant narratives of free-market globalization, open innovation, and digital development that ostensibly promise inclusivity and equality but often mask the perpetuation of extractive relationships. This study reveals how these narratives co-opt local adoption of emerging business models, innovations, and practices in developing economies while prescribing modes of organizing that align with the profit-maximizing imperatives of global capitalism (Couldry & Mejias, 2019). By situating AI development within the historical context of colonial enterprise, this research underscores how developing economies in global South, are positioned as open markets for finished technological products while simultaneously serving as sources of labor and entrepreneurial talent for mobilizing borrowed technologies for local contexts. 
 
This research employs critical discourse analysis of texts, media reports, and corporate narratives surrounding TikTok and Huawei, two Chinese corporations that have challenged the dominance of Western firms. Drawing on various media sources (NYT, Economist, and Financial Times), as well as scholarly critiques of digital capitalism (Schiller, 1999), the study examines the systematic vilification of these firms. This discourse, it is argued, is not solely an economic maneuver to protect Western market dominance but also a geopolitical strategy to neutralize potential shifts in the global centers of accumulation, extraction, and profitability. The case of India, where TikTok was banned under the pretext of national security amid ongoing border tensions, exemplifies how such measures are often employed to entrench the dominance of American platforms like YouTube, Instagram, and Twitter, which now monopolize India’s social information and communication market (Bhattacharya, 2021). Even when firms from alternative locations demonstrate capacity to compete on innovation, quality, and accessibility, e.g. (TikTok and Huawei), they face intervention through sovereign state concerns, such as national security, that align with broader Western imperial interests (Singh & Chander, 2022). Furthermore, the debates on data colonialism, global governance, and AI ethics are shown to reinforce, rather than disrupt, these dynamics. Concepts of inclusivity, diversity, and equality are often co-opted to create a veneer of legitimacy while preserving the core-periphery structures of global capitalism.
 
 
The results of this research reveal that the global technology ecosystem is a deeply hierarchical structure where access to markets, resources, and legitimacy is mediated by Western-centric norms and geopolitical considerations. 

CO-AUTHORS
—
ABSTRACT

In this paper we have dwelled into exploring out the interaction of AI with the electoral process, considering this is the key component of democracy, it becomes pertinent to understand the exploitative nature of AI which is being misused to undermine the democratic principles. We have chosen the jurisdictions of United States of America, European Union and India for this paper. It Is well documented that there has been significant enthusiasm across the globe in respect of using AI for all forms of social activity. However, the electoral process—the time, place, and manner of elections within democratic nations—is one of few sectors in which there has been limited penetration of AI. Electoral management bodies in many countries have recently started exploring and deliberating over the use of AI in the electoral process.  Artificial Intelligence (AI) has emerged as a transformative force in political campaigns and elections, significantly influencing freedom of expression through algorithmic content curation, micro-targeting, and automated decision-making systems. While AI has the potential to enhance democratic participation, its unchecked deployment poses serious risks to political discourse, media pluralism, and electoral integrity. The absence of transparency and accountability mechanisms in AI-driven content moderation and promotion has led to the amplification of disinformation, suppression of marginalized voices, and manipulation of voter perceptions.

The self-regulation model adopted by digital platforms has proven inadequate in addressing these challenges, necessitating a robust regulatory framework to ensure AI operates in alignment with human rights principles. The European Commission’s proposed Digital Services Act (DSA) and Digital Markets Act (DMA) mark an important step toward regulating online platforms, but additional measures are required to ensure transparency, accountability, and fairness in AI-driven political communication. A comprehensive legal framework should mandate independent audits of AI systems, public access to data for research purposes, and standardized content moderation guidelines developed in collaboration with civil society organizations.

The democratization of AI is imperative to prevent monopolization by a few tech conglomerates and ensure equitable access to AI tools for political actors, journalists, and civil society. Open-source AI solutions, interoperability standards for content moderation, and public oversight mechanisms should be prioritized to counterbalance the power of private entities in shaping political narratives. Additionally, the regulation of micro-targeting and digital political advertising must be strengthened to prevent manipulation and ensure transparency in electoral processes. A global regulatory approach, rooted in international human rights standards, is needed to address the cross-border nature of AI-driven political influence. Governments, policymakers, and tech companies must collaborate to create a governance model that upholds democratic values, safeguards freedom of expression, and promotes an open, fair, and inclusive digital space. Without proactive intervention, AI risks becoming a tool for digital authoritarianism rather than a force for democratic empowerment.

In this paper, we consider five avenues within the core electoral process which have potential for AI usage, and map the challenges involved in using AI within them. These five avenues are: voter list maintenance, determining polling booth locations, polling booth protection processes, voter authentication, and video monitoring of elections. Within each avenue, we lay down the context, illustrate current or potential usage of AI, and discuss extant or potential ramifications of AI usage, as well as potential directions for mitigating risks when considering AI usage. We believe that the scant current usage of AI within electoral processes provides a very rare opportunity to deliberate on the risks and mitigation possibilities prior to actual and widespread AI deployment. This paper is an attempt to map the horizons of risks and opportunities in using AI within electoral processes and to help shape the debate around the topic.

 

Keywords: Artificial Intelligence, Elections, Corrupt Practices, Democracy, Electoral Process

CO-AUTHORS
—
ABSTRACT

Artificial Intelligence (AI) is transforming public policy formulation by making governance more participatory, transparent, and responsive. In the digital age, social media platforms serve as an open forum for public discourse, where citizens engage in policy discussions, critique government actions, and advocate for change.Traditional policymaking has often been a top-down process, where decisions are driven by bureaucratic and political considerations, often ignoring the voices of marginalized communities. AI, through natural language processing (NLP) and sentiment analysis, enables governments to analyze large-scale social media interactions, extract key themes, and assess public sentiment in real time. This data-driven approach ensures that policies are aligned with public concerns rather than elite preferences, making AI a critical tool for democratising governance.

 

A fundamental component of AI-driven policy formulation is text as “communicative rationality”, where public discussion and reasoning contribute to governance decisions. Drawing from Jürgen Habermas' theory of communicative action, this approach emphasises that rational discourse, rather than political coercion, should shape public policies. AI facilitates this process by analysing millions of comments, petitions, and discussions across digital platforms, allowing policymakers to identify dominant narratives, public grievances, and social priorities. By translating digital discourse into actionable insights, AI ensures that governance becomes more deliberative and inclusive. Governments can now use AI-powered discourse mapping to identify policy gaps, understand the effectiveness of existing measures, and predict public reactions before policy implementation.

 

The objective of this study is to explore how AI-driven analysis of social media discourse enhances participatory governance and contributes to a more democratic public policy approach. To achieve this, I will examine how AI processes public discussions using NLP-based sentiment analysis and opinion mining, providing a quantitative and qualitative understanding of policy-related conversations. Additionally, I will analyze real-world case studies where AI-driven public sentiment analysis influenced governance decisions in areas like development and displacement, political representation and recruitment , and social welfare programs. Furthermore, I will investigate how AI-powered feedback loops enable continuous policy adaptation, ensuring that public policies remain dynamic and reflective of evolving citizen needs.

 

While AI presents vast opportunities for democratic governance, challenges such as algorithmic bias, misinformation, and ethical concerns must be addressed to prevent AI-driven policymaking from reinforcing existing inequalities. This paper will also explore frameworks for responsible AI governance, ensuring that AI remains an enabler of democracy rather than a tool for surveillance or manipulation. As public discourse continues to shift to digital platforms, AI-powered policymaking will serve as a bridge between citizens and governments, fostering a governance model that is more inclusive, responsive, and democratic.

 

CO-AUTHORS
Paridhi Sharma (OP Jindal Global University, Sonipat )
ABSTRACT

Towards Responsible AI in India: A Strategic Framework for Robust Governance and  Impact Assessment of AI Systems

This paper will delve into the transformative impact of Artificial Intelligence (AI) across various sectors while highlighting the ethical, legal, and social challenges arising from its rapid advancement. AI has demonstrated immense potential in revolutionising industries, enhancing efficiency, and contributing to societal progress. However, these benefits also raise significant concerns about fairness, accountability, transparency, privacy, security, and adherence to human-centric values. Addressing these issues is crucial to ensuring that AI development and deployment align with the principles of Responsible AI.

To facilitate the responsible governance of AI technologies, it is imperative to establish a comprehensive and structured "AI Impact Assessment Mechanism" tailored to the Indian context. This mechanism would serve as a regulatory and evaluative framework to systematically assess AI systems and applications, ensuring they conform to ethical standards, legal mandates, and policy guidelines. By implementing such an assessment mechanism, policymakers and regulatory authorities can gain a thorough understanding of the AI ecosystem in India. This will aid in mitigating potential risks while maximising AI's benefits in a controlled and accountable manner.

In recent years, India has taken significant steps toward AI governance, particularly through the formulation of the National Strategy on AI (2018) and its adherence to Responsible AI principles endorsed by global organizations such as the Organisation for Economic Co-operation and Development (OECD) and the United Nations Educational, Scientific and Cultural Organization (UNESCO). These policy frameworks emphasize the necessity of fairness, inclusivity, and ethical considerations in AI applications. Despite these efforts, the absence of a structured impact assessment mechanism leaves a regulatory gap that must be addressed to ensure AI technologies are deployed in a manner that aligns with national and international best practices.

The proposed "AI Impact Assessment Mechanism" aims to bridge this gap by integrating a multi-dimensional evaluation process that scrutinizes AI applications through legal, ethical, and societal lenses. The mechanism would involve key stakeholders, including government agencies, AI developers, industry experts, and civil society organizations, to create a balanced and inclusive approach toward AI governance. This framework would facilitate risk assessment, regulatory compliance checks, and continuous monitoring of AI systems to ensure they adhere to India's existing legal framework and ethical norms.

Additionally, this assessment model would empower the government with a structured oversight mechanism that fosters trust and accountability in AI deployment. By enforcing compliance with ethical AI principles, it would contribute to reducing bias, preventing misuse, and promoting transparency in AI-driven decision-making processes. The AI Impact Assessment Mechanism would also enhance public confidence in AI technologies by ensuring their implementation prioritizes social well-being and adheres to human rights standards.

 

Keywords: - Artificial Intelligence; Assessment; Co-operation and Development (OECD); United Nations Educational, Scientific and Cultural Organization (UNESCO)

 

CO-AUTHORS
—
ABSTRACT

This research investigates strategies to promote the participation of Mauritian women in Artificial Intelligence (AI) within the context of the country’s emerging AI landscape. The primary research question is: What is Mauritius doing to promote the participation of Mauritian women in AI? Despite the growing momentum of AI development in Mauritius, there is limited research on the participation of women in the sector.

 

The study breaks down the main question into four sub-questions: (1) What is the status of AI in Mauritius (with a special focus on gender)? It assesses the current AI landscape in Mauritius, while exploring gender disparities within the sector. It probes whether women face marginalisation or if AI-related incentives are offered. It evaluates whether Mauritian women have equal opportunities as their male counterparts to rise as leaders in AI development and implementation.

 

(2) What challenges do Mauritian women face in pursuing careers in AI? It evaluates the challenges faced by secondary schools in guiding and supporting girls towards AI career paths and assesses the barriers female students encounter at the tertiary level. It also evaluates how well the work environment supports women in AI-related fields.

 

(3) What strategies currently exist to promote their participation? Investigating existing policies and initiatives will help evaluate current efforts and identify areas for improvement. (4) What additional measures can be adopted to enhance their engagement? It examines measures to be taken or enforced at the secondary school level to guide girls toward AI careers, at the tertiary level to encourage women to pursue AI, workplace initiatives to support women in AI, and collaborative policies between the government and private sector to promote women’s participation in AI.

 

A concurrent mixed-methods approach is employed, combining quantitative data from 80 professionals in the ICT and AI sectors with qualitative insights from secondary research. Statistical analysis using SPSS, including descriptive statistics, t-tests for gender differences as well as correlation analyses, would be conducted. The qualitative component involves thematic analysis of existing AI-related policies and academic literature.

 

Gender is crucial to democratization. This research aligns with the panel theme Democratising Artificial Intelligence: Putting People at the Centre and addresses the panel’s research question: How can regulatory frameworks be designed to proactively manage the societal risks and benefits of AI-based technologies to ensure equitable outcomes? By focusing on Mauritius, the first African country to adopt a national AI policy, this study provides insights into how gender-inclusive AI frameworks can be developed in the African context. Notably, despite Mauritius being a leader in Africa's IT sector, the National AI Strategy and the Digital Mauritius 2030 Strategic Plan, which guide AI development, both lack specific measures to promote women’s participation in AI.

 

Additionally, the World Economic Forum (2023) highlights that women are mostly in the entry-level roles in STEM fields, which is why this panel has been chosen – to emphasise the need for women to be at the center of AI development in democratic governance.

 

It should be noted that a substantial amount of research and data have already been collected for this study and just needs to be finalised into a paper.

CO-AUTHORS
Aditya Agrawal (OP Jindal Global University)
ABSTRACT

With Large Language Models or LLMs and Generative AI, such as ChatGPT and Deepseek, taking over the world, the question of internet/AI governance becomes vital. Since AI models are trained on existing literature, they are implicitly and unsuspectingly trained on biased material creating critical issues of gender imbalance. Therefore, in order to have an inclusive and equitable technological system, it becomes important to accept that there exist an implicit or unintentional bias. To understand the tech gender bias, the translation from one language to another of gender neutral terms is noteworthy. The results will show that the Language processing models are associating the certain professions, behaviours or roles with specific genders. The example reinforces the programmers’ biases or skewed patterns inflicted on machines while feeding training data. The huge diversity in the answers to series of sensitive geo-political questions posted to both famous ChatGPT and newly launched Deepseek by the news platform inshorts under insights by inshorts on January 31, 2025 is yet another recent depiction of transfer of human biases to machine through data transfer. Shifting from the abstract theoretical implementation of AI, the paper is advocating inclusive, equitable and diverse approach for AI creation and implementation. With feminist analysis of current AI policies the research seeks to redefine the intersection of technology, gender and society. While highlighting the advent of school of Feminist Artificial Intelligence the paper critically explore the historical exclusion of women’s knowledge and perspective in scientific discourse. Promoting Alison Adam’s critique of AI, the authors doesn’t promote FAI as an alternate but as a means to well-informed and inclusive AI. Signifying the trajectory and progression for an inclusive AI, the research focuses on importance of Models, design and data collection that can be very well envisaged in progression of conceptual framework to tangible applications.

 

The research centrally argue for development of AI technologies that stands as a liberating force representing feminist theory. With AI slowly seeping into all the major sectors around the world from healthcare to legal aid to pharma to food industry, it’s imperative to calculate the cost of tech discrimination to society as whole. The authors realising the risk associated with such biases provides a three pronged feminist approach to the AI development and implementation. This research through a performative model of awareness, inclusion and regulation stresses for a more democratic policy making to bridge the gendered gap. Being cognizant of the importance of broader initiatives and collective solidarity to mitigate such discrimination the research suggests mandatory tech auditing and sensitization to have an inclusive AI transformation.

CO-AUTHORS
Aditya Agrawal (OP Jindal Global University)
ABSTRACT

Artificial Intelligence (AI) has rapidly evolved from a futuristic concept into an indispensable tool in various industries, reshaping economies, governance, and everyday life. As its influence expands, prominent figures in the technology sector, including Sundar Pichai, Elon Musk, and Brad Smith, have called for proactive regulation to mitigate its risks. AI-powered applications such as deep fakes, facial recognition, and autonomous systems pose significant threats to public safety, privacy, and democratic institutions. The debate over AI regulation is multifaceted, encompassing ethical, legal, and geopolitical concerns.

A central question in the AI discourse is whether AI should be regulated at all, and if so, who should oversee its governance? While proponents of regulation argue that unchecked AI development could lead to existential threats—including the misuse of autonomous weapons and AI-driven disinformation campaigns—opponents caution against premature regulatory intervention that could stifle innovation. Azamat Abdoullaev contends that AI, as a fundamental technology, should remain free from early-stage regulations, as improper oversight may hinder its growth. However, figures like Stephen Hawking and Bill Gates warn that an unregulated AI landscape could lead to disastrous consequences, emphasizing the urgent need for ethical and legal safeguards. The regulatory dilemma presents two contrasting approaches: state-imposed regulation versus self-regulation by industry players. Government-led regulations offer a structured framework to ensure accountability and fairness, but excessive control may impede technological progress and lead to uneven global competitiveness. In contrast, self-regulation relies on corporate responsibility, which, while fostering innovation, risks creating an environment where ethical concerns are overlooked in the pursuit of market dominance. A balanced approach, incorporating mandatory ethical standards and sector-specific oversight, is crucial to navigating AI governance effectively.

India's stance on AI regulation, as outlined in NITI Aayog's ‘National Strategy for Artificial Intelligence’ and the 2020 Working Document on Responsible AI, underscores the need for a flexible, risk-based regulatory framework. The proposed oversight body, comprising experts from various disciplines, aims to develop sector-specific guidelines, ensuring AI deployment aligns with national interests and ethical considerations. However, the absence of a comprehensive AI-specific legal framework in India, as noted by legal scholars, highlights gaps in policy execution and enforcement. Internationally, efforts such as the European Union's Digital Services Act (DSA) and Digital Markets Act (DMA)exemplify legislative attempts to regulate AI-driven platforms, ensuring accountability, transparency, and user rights protection. However, the lack of global consensus remains a significant hurdle, as divergent regulatory frameworks across countries create inconsistencies in AI governance.

This research paper will examine the need for AI regulation, ethical dilemmas, privacy concerns, and security threats posed by unregulated AI. Secondly we have discussed the key stakeholder perspectives, analysis of views from technology leaders, policymakers, legal scholars, and AI ethicists. We have also discussed the global regulatory approaches for which we have included case studies from the EU, the US, and India, exploring existing and proposed AI governance models. Authors have also tried to address the challenges in AI regulation and have attempted to provide the future roadmap for AI governance for a globally coordinated regulatory framework, emphasizing transparency, accountability, and ethical AI development.

 

Keywords: Artificial Intelligence, Humanised AI, Democracy, Public Policy, Responsible AI.

CO-AUTHORS
Anupam Guha (Indian Institute of Technology, Bombay)
ABSTRACT

As Artificial Intelligence (AI) rapidly transforms various sectors, from finance, healthcare, and government functions to transportation, manufacturing, and education, governments worldwide are developing policies for its adoption. These policies are crucial in determining how society adopts AI technology (Guenduez & Mettler, 2023). Other stakeholders, including private companies, media, research institutions, and think tanks, also significantly shape the AI policy landscape (Deshpande & Sharp, 2022). The stakeholders often influence policy processes through narratives and agenda-setting. 

Hayes & Fitzpatrick (2024) argue that narratives inform and structure relations between people and technology, and dominant groups are better positioned to control these narratives. Numerous stakeholders, including tech companies and management consultancies, regularly employ sophisticated public relations campaigns to position AI as an indispensable and trustable part of the future using whitepapers, surveys, and reports (Bourne, 2019).

The Narrative Policy Framework (NPF) is a theoretical approach which analyses how policy narratives shape public opinion and policy outcomes. NPF is a useful method for understanding stories that support and reinforce the norms and rules of institutions. The Framework asserts that "policy narratives have some combination of a setting, characters (heroes, victims, and villains), plots, and a moral of the story (policy solution)." These are the four "key elements" of narratives (Jones et al., 2014). In the context of AI policy, NPF can help identify and analyse the narratives constructed by the stakeholders to understand their impact on AI governance. This technique lends itself to analysis by NLP tools, historically used for narrative analysis. In particular, temporal topic modelling like Bayesian hidden topic Markov model (Gruber et al., 2007) and Relationship Modeling Networks (Iyyer et al., 2016) have been used to analyse character relationships in stories. 

Our work uses these NLP methods combined with qualitative discourse analysis to analyse government documents, policy papers, statements, and reports from private companies, newspaper articles, reports, and research publications. Furthermore, this research clarifies the scope, significance, and consequences of using machine learning in social research. Expanding the scope of research in these directions can provide a more nuanced understanding of AI governance and the role of stakeholders in shaping public opinion and policy.

CO-AUTHORS
—
ABSTRACT

The pervasiveness and ubiquity of AI has drawn attention to the potential dangers and threats that algorithmic systems pose in different parts of society. Artificial intelligence disruptively transforms decision-making and the performance of various tasks in society. A rigorous body of literature draws attention to potential threats such as injustice, discrimination, exclusion, and various organizational risks. In response to these AI threats, the literature places democracy as essential. An emerging body of literature has constructed claims for AI democratization, without precisely defining what the meaning of democracy is with which we seek democratization. The common thread in this emerging struggle is that democratizing AI means expanding society’s access to tools, data sets, algorithms, and computing power. In this article, we argue that the meaning of AI democratization goes beyond this issue. The problem with AI democratization is how we institutionalize the decision-making process involving complex interactions between humans and artificial agents. In this sense, AI democratization must be thought of within emerging algorithmic regimes. The objective of this article is to analyze the concept of AI democratization and to consider possible frameworks for the insertion of AI in the dynamics of democratic political regimes. 

CO-AUTHORS
Sushant Kumar (Jindal School of Government and Public Policy)
ABSTRACT

Across the world  biometric systems are increasingly being adopted by States like India, Nigeria, United States and UAE,for more ‘secure, accurate, efficient’ governance. In the domain of education administration as well, biometric systems have several use cases ranging from monitoring teacher attendance to ascertaining ‘real’ numbers in terms of student enrolment by weeding out false or duplicate identity claims. In this paper I will go over prominent use cases across the world, proceeding to then theorise both the discourse they are embedded in as well as their materiality using concepts from Science and Technology Studies.

Science and Technology studies serves as an especially fruitful conceptual toolbox to navigate understanding of these processes. I will tracing how the idea of ‘truth’ is attached to the informationalised body and then this bio-information isused as a state sanctioned sorting deviceto distribute benefits/rewards to the ‘deserving’. Infrastructure surrounding the public delivery of services is built on the affordance of certainty and the claim of immunity to proxies this technology provides.

It is also useful to view it as an obligatory passing point - both for the state and the citizen, alongside tech companies, civil society and other intermediaries who form an assemblage to produce and stabilise this infrastructure of truth making.In addition, each actor in this assemblage of the socio-technical artifact has diverse meanings and narratives about it, some overlapping some discrete, some even contradictory. Such socio-technical artifacts embody state making projects, holding a special significance for developing and postcolonial societies. In the latter contexts there is a tendency to slide into uncritical triumphalism with regard to participation in modernity through technological adoption. Therefore, in these paper I will also critically read these narratives to evaluate whose perspectives are adopted and whose get marginalised, querying for questions of state surveillance, privacy and exclusion.