Format

NR50. Human Futures of AI: Foresight and Public Policy

Research
PANEL CHAIR(S)
Preeti Raghunath
University of Sheffield
Main chair
CATEGORISATION
POLICY TOPIC
Research
SECTOR
—
KEYWORDS
—
GENERAL OBJECTIVES, RESEARCH QUESTIONS AND SCIENTIFIC RELEVANCE

This panel seeks to bring together scholars and practitioners focused on developments in forms of Artificial Intelligence, and corresponding and adjacent Public Policies. In doing so, the panel will create space to engage with AI futurisms, practical methods to explore building foresight knowledge and pathways for near-futures, and engaged and ethnographic research. The objective of this panel is to centre Human Futures in times of AI, and should be of interest to both scholars and practitioners working at the intersections of Time, Policy and Technology, broadly put. 



The Research Questions that papers on this panel could explore are:

 

1. How does futures thinking manifest in policymaking for AI? Who futures are these? What values and ideologies are championed in the process?

 

2. To what extent and how do AI policies and their futures-orientation impinge upon, interact with and shape (or not) sectoral public policies? How does this change the ways in which these sectors shape human lives, livelihoods and opportunities? 




This panel welcomes a variety of formats, including and beyond traditional papers, such as experimental thinking, creative-academic engagements, foresight and strategic position papers, and more. Authors are particularly encouraged to go beyond binaries of AI as good/bad, and help us expand our understanding by unravelling the complexities of human lives that interface with AI.

 

CALL FOR PAPERS

Globally, surveys of citizens and members of the public reveal the lack of trust in what are increasingly AI-embedded public-private products and services across key sectors. While there is growing recognition of the benefits of AI, with medical diagnostics, improved access to learning materials, language and translation or automation of mundane tasks, there’s a lot of skepticism as well. These concerns are around unanticipated and hidden risks, especially around time-bound imaginaries, such as near- and long-termism and risks associated with each of them. There also exists a huge evidence gap in that we know very little about how citizens/users/consumers across contexts and settings approach and use AI. We do not know much about peoples’ aspirations for AI, why resistance exists today and how they’d respond in future risk scenarios.

 

 Foresight and behavioural methods offer insights on behaviours, needs and capabilities in confronting risks, and building capacities for Safer AI. This panel, therefore, aims to bring together research and practices on a human-first approach to AI policy, where the focus is not on developing foresight for AI Futures, but for Human Futures amidst advanced AI.

 

The panel aims to centre “Lived policy”, which are “how public policies are lived by those impacted by them” (Ali, 2023) and Deliberative Policymaking (Raghunath, 2020) for AI. Towards this, this panel aims to facilitate conversations on foresight and futures approaches in combination with qualitative, ethnographic and behavioural methods, to collectively think about near-future lived experiences of diverse groups of people in a range of sectors. Drawing on a range of behavioural and futures methods (eg. rapid ethnographic assessments, scenarios-building, ethnographic experiential futures), this panel will explore themes at the intersections of Time, Policy and Technology.

 

This panel welcomes a variety of formats, including and beyond traditional papers, such as experimental thinking, creative-academic engagements, foresight and strategic position papers, and more. Authors are particularly encouraged to go beyond binaries of AI as good/bad, and help us expand our understanding by unravelling the complexities of human lives that interface with AI.



ACCEPTED
Format
01 - Decolonial Pathways to AI Personhood: Frameworks for Inclusive Futures
Sanchet Sharma (National Law School of India University)
CO-AUTHORS
—
ABSTRACT

This paper critically examines whether granting legal personhood to artificial intelligence (AI) systems could recreate colonial patterns of impunity. We draw on the history of colonial trading companies, whose limited liability and extraterritorial jurisdiction enabled exploitation, to show how AI personhood might likewise diffuse responsibility while concentrating power.

We explore this through a literature review of colonial trading companies and their social impacts, paired with an examination of documented failures in areas such as autonomous transport, automated welfare screening, and high-stakes medical diagnosis, including recent incidents in the Global South. Across these disparate cases, we find recurring mechanisms that scatter accountability while profit and control remain concentrated. The persistence of this pattern across centuries signals the need for a new accountability framework for AI personhood. Building on these findings, we propose a preliminary risk-tier model inspired by the EU AI Act. For a certain grade of AI-related risk, we suggest that a named human or firm should hold strict, non-transferable liability, ensuring responsibility rests with a real actor capable of addressing harm. We will be further reviewing this tiered structure as the paper develops.

ACCEPTED
Format
02 - Environment Cost of India's AI Boom
Escandita Tewari (Sustainable Futures Collaborative)
CO-AUTHORS
—
ABSTRACT

Artificial Intelligence as Climate-Exposed Infrastructure. Identifying Early System Failures in India’s AI Expansion

India’s push to become a global hub for artificial intelligence is being framed as a digital and economic leap. Yet the material systems that support AI are already under stress. This paper approaches India’s AI expansion through a systems lens that asks two questions. What breaks first? And where do policy blind spots allow fragility to scale?

 

AI growth depends on data centres, semiconductor fabrication and high-performance computing infrastructure. These systems are highly sensitive to electricity reliability, water availability and thermal conditions. In India they are increasingly being sited in regions facing heat stress, water scarcity and grid constraints. As AI workloads grow, these pressures compound. The first points of failure are not abstract climate risks but operational ones. Cooling systems under heat waves. Water sourcing during droughts. Power demand spikes that strain local grids. These failures translate into economic risk long before they appear as environmental externalities.

 

Current policy frameworks treat AI as an immaterial digital sector. Environmental regulation disclosure norms and industrial siting rules lag behind the physical footprint now emerging. This paper argues that without early intervention India risks locking in a resource-intensive AI ecosystem that is brittle under climate stress.

 

The paper highlights the policy gaps that exist in India's current AI governance and proposes a governance approach that treats AI infrastructure as critical climate-exposed infrastructure. It argues that aligning AI policy with climate resilience is not a constraint on growth. but a condition for durable growth.


 

ACCEPTED
Format
03 - Who Decides When No One Decides? Human Futures in Algorithmic Governance
Olivia Ruhil (National Centre for Good Governance, Government of India )
CO-AUTHORS
—
ABSTRACT

This chapter examines how the growing incorporation of algorithmic systems into public governance alters the institutional conditions under which judgment, responsibility, and collective memory are exercised. It advances the hypothesis that contemporary AI-driven infrastructures do not merely improve administrative efficiency; they reshape how law, regulation, and policy are encountered in everyday life, with consequences for how human futures are organised and governed.

 

The analysis focuses on policy frameworks structured around optimisation and automation, showing how these approaches embed particular assumptions about human behaviour and social order. As predictive models built from historical data increasingly guide decision-making, institutional practice shifts from interpretation toward repetition. Processes once reliant on contextual reasoning begin to privilege procedural consistency, narrowing the space available for dissent, exception, and institutional adaptation. In this setting, deviation tends to appear as risk, and policy outcomes often reproduce existing distributions of advantage rather than re-examine them.

 

Methodologically, the chapter draws on theoretical and institutional analysis of developments in algorithmic decision-making, predictive governance tools, and digital regulatory infrastructures. It brings together work on algorithmic judgment, memory, and authority to show how optimisation-driven systems convert past records into expectations about future conduct. At the same time, digital platforms shape what remains visible, credible, and actionable within governance processes, influencing whose claims are recognised and whose experiences remain peripheral.

 

The chapter further considers the consequences of generative systems capable of producing legal and administrative language that appears authoritative while remaining detached from identifiable processes of deliberation. In parallel, algorithmic coordination across markets and regulatory systems increasingly generates outcomes without explicit agreement or negotiation. Authority persists, yet responsibility becomes difficult to locate, complicating both democratic oversight and institutional learning.

 

Viewed through the lens of human futures, these developments suggest that policy imaginaries centred on innovation and automation risk reducing the scope for judgment, disagreement, and rehabilitation in public life. Policy is encountered not only through formal rules but through data-driven systems that shape eligibility, credibility, and exposure to risk, often beyond visible sites of contestation.

 

The chapter concludes by reframing foresight in AI policy as an institutional rather than solely technological question. The durability of governance depends upon preserving space for judgment, contestation, and repair within increasingly automated systems. Discretion, memory, and dissent remain essential to sustaining public authority that continues to answer to the lives it governs in periods of rapid technological change.

ACCEPTED
Format
04 - AI adoption and its effects on job outcomes in select sectors in India
Manav Khaire (Indian Institute of Technology Palakkad)
CO-AUTHORS
Pramath Kant (Infosys Ltd.)
ABSTRACT

India's orientation towards AI governance emphasizes a wait-and-watch approach, prioritizing the mitigation of the risk posed by AI-related harm. Notably, India's human-centric approach upholds the fundamental rights enshrined in the Constitution of India while guiding the basic principles of responsible AI, including safety, equality, inclusivity, non-discrimination, and privacy (NITI Aayog, 2025). This approach aligns with the European Union's AI governance framework, which emphasizes personal data protection, risk-based evaluation of AI systems, and human-centricity. Globally, AI policies, characterized by their future-oriented approach to ethics, innovation, and risk mitigation, increasingly intersect with labour-intensive industries such as financial services, telecom, and manufacturing, including automotive sector. These interactions occur through corporate advocacy for adaptive regulations that balance rapid AI deployment with compliance, thereby reshaping policy landscapes to accelerate automation of routine and labour-intensive tasks, with the potential to alter the lives, livelihoods, and opportunities of the labour class. In the US, the bipartisan AI-Related Job Impacts Clarity Act (2025) mandates quarterly disclosures by companies and agencies on AI-driven layoffs and retraining, highlighting transparency efforts even as it facilitates automation of routine tasks. The EU AI Act (2024) regulates high-risk workplace AI like emotion inference bans, prioritizing safety but enabling productivity tools that displace labour.

Telecom giants Bharti Airtel and Reliance Jio embed AI in network optimization and customer personalization under initiatives like India's National Technology Policy 2025, influencing spectrum policies and digital inclusion strategies; yet, this automates call centres and maintenance tasks, narrowing opportunities for low-skilled labour while generating specialized tech jobs. Manufacturing leaders such as Siemens drive AI adoption in predictive maintenance and smart factories, pushing for energy-efficient regulations and vocational training mandates; with over half of manufacturers already integrating AI, productivity surges but manual assembly lines face obsolescence, intensifying substitution effects on working-class employment.

These bidirectional dynamics—where sectoral policies inform AI guardrails and vice versa—fundamentally alter human lives, livelihoods, and opportunities. Even though companies and industry representatives recognize the transformational nature of AI, leading to new job creation, there is little evidence of how this will happen. More often, supporters of this argument fall back on the historical claim that the internet revolution created more jobs; however, their argument is merely a set of positive expectations, lacking empirical evidence. Our study investigates the impact of AI automation on jobs and the nature of the automation on the working class. Through secondary data analysis and stakeholder interviews, this study examines how the adoption of AI policies has affected the workforce across sectors, including telecom, manufacturing, financial services, and professional services. We argue that AI-led automation will erode traditional jobs for the working class, exacerbating inequality and possibly worsening social unrest. Yet it will also foster innovation-driven growth and upskilling pathways, requiring inclusive policies to democratize AI benefits. Ultimately, this convergence necessitates a profound discussion on the emerging "socio-AI-economic" paradigm, urging proactive governance for equitable futures.

ACCEPTED
Format
05 - The future of women’s work in the age of AI: A case study of the transformation of brick and mortar beauty salons to beauty work on platforms
Sreyan Chatterjee (The Manchester Metropolitan University)
CO-AUTHORS
—
ABSTRACT

Brick-and-mortar beauty salons are ubiquitous across the globe and so is the beauty work that is undertaken therein. While the nature of beauty standards were often seen as control mechanisms for extant patriarchal norms, more recent feminist studies of beauty work have characterised  these spaces being unique gendered spaces with potential of female agency to redefine and shape beauty norms in ways that work for women as well as open up opportunities for women to build cross-class/caste/race solidarities. Beauty platforms have secured significant market share in India - and platforms with relatively smaller footprints are emerging in West Asia, Southeast Asia as well as in the United Kingdom. 


Algorithmic management or AI-augmented human resource management (‘HRM’) provides a space away from physical surveillance of the brick-and-mortar salon and yet increases the organisational anxiety of the algorithmic HRM framework recruiting clients as hybrid agents of surveillance in order to triangulate the digital control exercised by the platform itself. Scholars have demonstrated for example in India that beauty salons provided a space for professionalism where societal norms can be suspended or diluted and the return to the client’s home harkens to historically oppressive caste relations. 


In this paper, the author proposes to undertake a study of transforming job crafting between brick and mortar salons and beauty work platforms. From traditional Taylorist techniques to digital Taylorism, technological change is introduced in the form of AI-augmented HRM to aid a different model of value extraction. Within the informal economy, the platform model works on a specific form of exploitation using speculative data markets and monopolising the data flows from the sector. The author argues that unlike other service work, beauty presents a special problem to the platform and AI workplace model - as gendered front-line service work which has a significant core of affective labour or emotional work. It is because of this nature of beauty work the platform model is unable to enter into many markets around the globe and where it indeed does, it has focused on the low-value, ‘base of the pyramid’ of beauty work. Beauty work is unique by having both workers and customers as female in most cases, and being a low barrier formal and informal work to many women. 


Methodologically, this paper is conceptual in nature, focussed on secondary empirical research to support the core arguments. The main frameworks that the authors engage with are labour process theory and organisational sensemaking. As a case study for future policymaking for AI-augmented workplaces, this paper offers insights in what the contours of future policy should focus on, particularly how the humanness of certain work can be valued accurately and protected adequately. Emotion work is seen as a core feature of humanness and the ability of future policy to regulate emotion work will be a key determinant of the quality of jobs that human beings (particularly women) do in AI-augmented workplaces such as on digital platforms and future iterations of platforms.

ACCEPTED
Format
06 - Between Design and Practice: AI, Non-Use, and Lived Policy in ASHAs Digital Work
Aanchal Singh (International Institute of Information Technology, Bangalore)
CO-AUTHORS
Carolyn Tauro (International Institute of Information Technology Bangalore)
Amit Prakash (IIIT-Bangalore)
ABSTRACT
Digital health tools are increasingly promoted as solutions for strengthening frontline public health systems, especially in low-resource settings. In India, growing partnerships among NGOs, government, and private actors have formed an ecosystem that introduces AI-based tools across different public health domains, including for frontline workers such as Accredited Social Health Activists (ASHAs). AI deployments in frontline health work are shaped by sociotechnical imaginaries, understood as collectively held visions of desirable futures that link technological progress with social order (Jasanoff & Kim, 2015, pp. 4–6). In public health, these imaginaries often emphasise efficiency, standardisation, and data-driven governance. However, this paper emphasizes whether these tools improve outcomes often defined differently by different actors while paying less attention to how frontline workers actually engage with, adapt, or ignore these technologies in everyday practice. 
 This paper examines the deployment of an AI-based chatbot “A-Sakhi,” (pseudonym), designed to provide standardised government guidelines to ASHA workers. Despite institutional backing, the chatbot experienced uneven and often limited use across both rural and urban settings. At the same time, our fieldwork shows that ASHAs in urban areas rapidly adopted other AI tools for a wide range of tasks, including planning MCHN session (Maternal, Child Health & Nutrition) and their  sector meetings, counselling, translation, resolving field-level doubts, and other work-related needs, often without formal training or governance guidance. These tools were also used for personal purposes, such as supporting children’s education and exam preparation, checking horoscopes, or simply engaging with the bot for fun and liesure. This contrast invites probe of technology non-use, partial use, and selective use as meaningful practices rather than simple failures of implementation. Non-use, in this context, should be understood as a deliberate, situated, and socially meaningful practice shaped by the conditions of everyday work (Wyatt, 2003; Satchell & Dourish, 2009; Baumer et al., 2015).
 
The study draws on a multi-location qualitative research design across rural and urban public health settings in Rajasthan, conducted in two phases . Methods include in-depth interviews with ASHAs, ASHA supervisors, Auxiliary Nurse Midwives (ANMs), programme administrators, NGO staff, funding partners, technology designers, and government officials, alongside observations of ASHAs daily work and focus group discussions with ASHAs and implementing partners. This approach allows us to situate the chatbot within the broader institutional, policy, and technological landscape shaping ASHAs work.
Lived policy scholarship shape how frontline actors interpret, negotiate, and rework policy in practice (Raghunathan et al., 2019; Raghunathan & Kalyanpur, 2021), and on sociotechnical imaginaries as visions of governance embedded in technology and institutions (Jasanoff & Kim, 2015). This paper treats ASHAs as active agents rather than passive implementers. Findings indicate ASHAs technology practices are shaped by changing programme requirements, performance targets, data reporting demands, limited connectivity, and the need to maintain trust with communities. While ASHAs are increasingly subject to digital monitoring and standardisation, they also make practical and ethical judgements about which tools to use, which to ignore, and how to repurpose technologies to support their work. Especially in urban settings, the non-use of the official chatbot alongside widespread use of other AI tools reflects a preference for flexible, multi-purpose technologies over narrowly designed institutional systems. 
We argue that digital health tools will only be effective if they are designed with serious attention to lived policy environments, sociotechnical imaginaries, and the meaningful practice of non-use. Rather than treating non-use as a problem to be fixed, policymakers and designers must recognise it as a form of agency that reveals misalignments between institutional designs and frontline realities. This study contributes to debates on responsible AI and digital governance by demonstrating how AI-enabled systems restructure frontline work while simultaneously depending on workers everyday work practices of interpretation, adaptation, and use within under-resourced, over-stretched public health systems (Bovens & Zouridis, 2002; Eubanks, 2018).
 
References
  • Baumer, E. P. S., Ames, M. G., Burrell, J., Brubaker, J. R., & Dourish, P. (2015). On the importance and implications of studying technology non-use. Interactions, 22(2), 52–56.
  • Bovens, M., & Zouridis, S. (2002). From street-level to system-level bureaucracies: How information and communication technology is transforming administrative discretion and constitutional control. Public Administration Review, 62(2), 174–184.
  • Eubanks, V. (2018). Automating inequality: How high-tech tools profile, police, and punish the poor. St. Martin’s Press.
  • Jasanoff, S., & Kim, S.-H. (2015). Dreamscapes of modernity: Sociotechnical imaginaries and the fabrication of power. University of Chicago Press.
  • Raghunathan, K., et al. (2019). 
  • Raghunathan, K., & Kalyanpur, N. (2021). 
  • Satchell, C., & Dourish, P. (2009). Beyond the user: Use and non-use in HCI. Proceedings of OZCHI.
  • Wyatt, S. (2003). Non-users also matter: The construction of users and non-users of the Internet. In N. Oudshoorn & T. Pinch (Eds.), How users matter (pp. 67–79). MIT Press.
CO-AUTHORS
—
ABSTRACT

This paper critically examines whether granting legal personhood to artificial intelligence (AI) systems could recreate colonial patterns of impunity. We draw on the history of colonial trading companies, whose limited liability and extraterritorial jurisdiction enabled exploitation, to show how AI personhood might likewise diffuse responsibility while concentrating power.

We explore this through a literature review of colonial trading companies and their social impacts, paired with an examination of documented failures in areas such as autonomous transport, automated welfare screening, and high-stakes medical diagnosis, including recent incidents in the Global South. Across these disparate cases, we find recurring mechanisms that scatter accountability while profit and control remain concentrated. The persistence of this pattern across centuries signals the need for a new accountability framework for AI personhood. Building on these findings, we propose a preliminary risk-tier model inspired by the EU AI Act. For a certain grade of AI-related risk, we suggest that a named human or firm should hold strict, non-transferable liability, ensuring responsibility rests with a real actor capable of addressing harm. We will be further reviewing this tiered structure as the paper develops.

CO-AUTHORS
—
ABSTRACT

Artificial Intelligence as Climate-Exposed Infrastructure. Identifying Early System Failures in India’s AI Expansion

India’s push to become a global hub for artificial intelligence is being framed as a digital and economic leap. Yet the material systems that support AI are already under stress. This paper approaches India’s AI expansion through a systems lens that asks two questions. What breaks first? And where do policy blind spots allow fragility to scale?

 

AI growth depends on data centres, semiconductor fabrication and high-performance computing infrastructure. These systems are highly sensitive to electricity reliability, water availability and thermal conditions. In India they are increasingly being sited in regions facing heat stress, water scarcity and grid constraints. As AI workloads grow, these pressures compound. The first points of failure are not abstract climate risks but operational ones. Cooling systems under heat waves. Water sourcing during droughts. Power demand spikes that strain local grids. These failures translate into economic risk long before they appear as environmental externalities.

 

Current policy frameworks treat AI as an immaterial digital sector. Environmental regulation disclosure norms and industrial siting rules lag behind the physical footprint now emerging. This paper argues that without early intervention India risks locking in a resource-intensive AI ecosystem that is brittle under climate stress.

 

The paper highlights the policy gaps that exist in India's current AI governance and proposes a governance approach that treats AI infrastructure as critical climate-exposed infrastructure. It argues that aligning AI policy with climate resilience is not a constraint on growth. but a condition for durable growth.


 

CO-AUTHORS
—
ABSTRACT

This chapter examines how the growing incorporation of algorithmic systems into public governance alters the institutional conditions under which judgment, responsibility, and collective memory are exercised. It advances the hypothesis that contemporary AI-driven infrastructures do not merely improve administrative efficiency; they reshape how law, regulation, and policy are encountered in everyday life, with consequences for how human futures are organised and governed.

 

The analysis focuses on policy frameworks structured around optimisation and automation, showing how these approaches embed particular assumptions about human behaviour and social order. As predictive models built from historical data increasingly guide decision-making, institutional practice shifts from interpretation toward repetition. Processes once reliant on contextual reasoning begin to privilege procedural consistency, narrowing the space available for dissent, exception, and institutional adaptation. In this setting, deviation tends to appear as risk, and policy outcomes often reproduce existing distributions of advantage rather than re-examine them.

 

Methodologically, the chapter draws on theoretical and institutional analysis of developments in algorithmic decision-making, predictive governance tools, and digital regulatory infrastructures. It brings together work on algorithmic judgment, memory, and authority to show how optimisation-driven systems convert past records into expectations about future conduct. At the same time, digital platforms shape what remains visible, credible, and actionable within governance processes, influencing whose claims are recognised and whose experiences remain peripheral.

 

The chapter further considers the consequences of generative systems capable of producing legal and administrative language that appears authoritative while remaining detached from identifiable processes of deliberation. In parallel, algorithmic coordination across markets and regulatory systems increasingly generates outcomes without explicit agreement or negotiation. Authority persists, yet responsibility becomes difficult to locate, complicating both democratic oversight and institutional learning.

 

Viewed through the lens of human futures, these developments suggest that policy imaginaries centred on innovation and automation risk reducing the scope for judgment, disagreement, and rehabilitation in public life. Policy is encountered not only through formal rules but through data-driven systems that shape eligibility, credibility, and exposure to risk, often beyond visible sites of contestation.

 

The chapter concludes by reframing foresight in AI policy as an institutional rather than solely technological question. The durability of governance depends upon preserving space for judgment, contestation, and repair within increasingly automated systems. Discretion, memory, and dissent remain essential to sustaining public authority that continues to answer to the lives it governs in periods of rapid technological change.

CO-AUTHORS
Pramath Kant (Infosys Ltd.)
ABSTRACT

India's orientation towards AI governance emphasizes a wait-and-watch approach, prioritizing the mitigation of the risk posed by AI-related harm. Notably, India's human-centric approach upholds the fundamental rights enshrined in the Constitution of India while guiding the basic principles of responsible AI, including safety, equality, inclusivity, non-discrimination, and privacy (NITI Aayog, 2025). This approach aligns with the European Union's AI governance framework, which emphasizes personal data protection, risk-based evaluation of AI systems, and human-centricity. Globally, AI policies, characterized by their future-oriented approach to ethics, innovation, and risk mitigation, increasingly intersect with labour-intensive industries such as financial services, telecom, and manufacturing, including automotive sector. These interactions occur through corporate advocacy for adaptive regulations that balance rapid AI deployment with compliance, thereby reshaping policy landscapes to accelerate automation of routine and labour-intensive tasks, with the potential to alter the lives, livelihoods, and opportunities of the labour class. In the US, the bipartisan AI-Related Job Impacts Clarity Act (2025) mandates quarterly disclosures by companies and agencies on AI-driven layoffs and retraining, highlighting transparency efforts even as it facilitates automation of routine tasks. The EU AI Act (2024) regulates high-risk workplace AI like emotion inference bans, prioritizing safety but enabling productivity tools that displace labour.

Telecom giants Bharti Airtel and Reliance Jio embed AI in network optimization and customer personalization under initiatives like India's National Technology Policy 2025, influencing spectrum policies and digital inclusion strategies; yet, this automates call centres and maintenance tasks, narrowing opportunities for low-skilled labour while generating specialized tech jobs. Manufacturing leaders such as Siemens drive AI adoption in predictive maintenance and smart factories, pushing for energy-efficient regulations and vocational training mandates; with over half of manufacturers already integrating AI, productivity surges but manual assembly lines face obsolescence, intensifying substitution effects on working-class employment.

These bidirectional dynamics—where sectoral policies inform AI guardrails and vice versa—fundamentally alter human lives, livelihoods, and opportunities. Even though companies and industry representatives recognize the transformational nature of AI, leading to new job creation, there is little evidence of how this will happen. More often, supporters of this argument fall back on the historical claim that the internet revolution created more jobs; however, their argument is merely a set of positive expectations, lacking empirical evidence. Our study investigates the impact of AI automation on jobs and the nature of the automation on the working class. Through secondary data analysis and stakeholder interviews, this study examines how the adoption of AI policies has affected the workforce across sectors, including telecom, manufacturing, financial services, and professional services. We argue that AI-led automation will erode traditional jobs for the working class, exacerbating inequality and possibly worsening social unrest. Yet it will also foster innovation-driven growth and upskilling pathways, requiring inclusive policies to democratize AI benefits. Ultimately, this convergence necessitates a profound discussion on the emerging "socio-AI-economic" paradigm, urging proactive governance for equitable futures.

CO-AUTHORS
—
ABSTRACT

Brick-and-mortar beauty salons are ubiquitous across the globe and so is the beauty work that is undertaken therein. While the nature of beauty standards were often seen as control mechanisms for extant patriarchal norms, more recent feminist studies of beauty work have characterised  these spaces being unique gendered spaces with potential of female agency to redefine and shape beauty norms in ways that work for women as well as open up opportunities for women to build cross-class/caste/race solidarities. Beauty platforms have secured significant market share in India - and platforms with relatively smaller footprints are emerging in West Asia, Southeast Asia as well as in the United Kingdom. 


Algorithmic management or AI-augmented human resource management (‘HRM’) provides a space away from physical surveillance of the brick-and-mortar salon and yet increases the organisational anxiety of the algorithmic HRM framework recruiting clients as hybrid agents of surveillance in order to triangulate the digital control exercised by the platform itself. Scholars have demonstrated for example in India that beauty salons provided a space for professionalism where societal norms can be suspended or diluted and the return to the client’s home harkens to historically oppressive caste relations. 


In this paper, the author proposes to undertake a study of transforming job crafting between brick and mortar salons and beauty work platforms. From traditional Taylorist techniques to digital Taylorism, technological change is introduced in the form of AI-augmented HRM to aid a different model of value extraction. Within the informal economy, the platform model works on a specific form of exploitation using speculative data markets and monopolising the data flows from the sector. The author argues that unlike other service work, beauty presents a special problem to the platform and AI workplace model - as gendered front-line service work which has a significant core of affective labour or emotional work. It is because of this nature of beauty work the platform model is unable to enter into many markets around the globe and where it indeed does, it has focused on the low-value, ‘base of the pyramid’ of beauty work. Beauty work is unique by having both workers and customers as female in most cases, and being a low barrier formal and informal work to many women. 


Methodologically, this paper is conceptual in nature, focussed on secondary empirical research to support the core arguments. The main frameworks that the authors engage with are labour process theory and organisational sensemaking. As a case study for future policymaking for AI-augmented workplaces, this paper offers insights in what the contours of future policy should focus on, particularly how the humanness of certain work can be valued accurately and protected adequately. Emotion work is seen as a core feature of humanness and the ability of future policy to regulate emotion work will be a key determinant of the quality of jobs that human beings (particularly women) do in AI-augmented workplaces such as on digital platforms and future iterations of platforms.

CO-AUTHORS
Carolyn Tauro (International Institute of Information Technology Bangalore)
Amit Prakash (IIIT-Bangalore)
ABSTRACT
Digital health tools are increasingly promoted as solutions for strengthening frontline public health systems, especially in low-resource settings. In India, growing partnerships among NGOs, government, and private actors have formed an ecosystem that introduces AI-based tools across different public health domains, including for frontline workers such as Accredited Social Health Activists (ASHAs). AI deployments in frontline health work are shaped by sociotechnical imaginaries, understood as collectively held visions of desirable futures that link technological progress with social order (Jasanoff & Kim, 2015, pp. 4–6). In public health, these imaginaries often emphasise efficiency, standardisation, and data-driven governance. However, this paper emphasizes whether these tools improve outcomes often defined differently by different actors while paying less attention to how frontline workers actually engage with, adapt, or ignore these technologies in everyday practice. 
 This paper examines the deployment of an AI-based chatbot “A-Sakhi,” (pseudonym), designed to provide standardised government guidelines to ASHA workers. Despite institutional backing, the chatbot experienced uneven and often limited use across both rural and urban settings. At the same time, our fieldwork shows that ASHAs in urban areas rapidly adopted other AI tools for a wide range of tasks, including planning MCHN session (Maternal, Child Health & Nutrition) and their  sector meetings, counselling, translation, resolving field-level doubts, and other work-related needs, often without formal training or governance guidance. These tools were also used for personal purposes, such as supporting children’s education and exam preparation, checking horoscopes, or simply engaging with the bot for fun and liesure. This contrast invites probe of technology non-use, partial use, and selective use as meaningful practices rather than simple failures of implementation. Non-use, in this context, should be understood as a deliberate, situated, and socially meaningful practice shaped by the conditions of everyday work (Wyatt, 2003; Satchell & Dourish, 2009; Baumer et al., 2015).
 
The study draws on a multi-location qualitative research design across rural and urban public health settings in Rajasthan, conducted in two phases . Methods include in-depth interviews with ASHAs, ASHA supervisors, Auxiliary Nurse Midwives (ANMs), programme administrators, NGO staff, funding partners, technology designers, and government officials, alongside observations of ASHAs daily work and focus group discussions with ASHAs and implementing partners. This approach allows us to situate the chatbot within the broader institutional, policy, and technological landscape shaping ASHAs work.
Lived policy scholarship shape how frontline actors interpret, negotiate, and rework policy in practice (Raghunathan et al., 2019; Raghunathan & Kalyanpur, 2021), and on sociotechnical imaginaries as visions of governance embedded in technology and institutions (Jasanoff & Kim, 2015). This paper treats ASHAs as active agents rather than passive implementers. Findings indicate ASHAs technology practices are shaped by changing programme requirements, performance targets, data reporting demands, limited connectivity, and the need to maintain trust with communities. While ASHAs are increasingly subject to digital monitoring and standardisation, they also make practical and ethical judgements about which tools to use, which to ignore, and how to repurpose technologies to support their work. Especially in urban settings, the non-use of the official chatbot alongside widespread use of other AI tools reflects a preference for flexible, multi-purpose technologies over narrowly designed institutional systems. 
We argue that digital health tools will only be effective if they are designed with serious attention to lived policy environments, sociotechnical imaginaries, and the meaningful practice of non-use. Rather than treating non-use as a problem to be fixed, policymakers and designers must recognise it as a form of agency that reveals misalignments between institutional designs and frontline realities. This study contributes to debates on responsible AI and digital governance by demonstrating how AI-enabled systems restructure frontline work while simultaneously depending on workers everyday work practices of interpretation, adaptation, and use within under-resourced, over-stretched public health systems (Bovens & Zouridis, 2002; Eubanks, 2018).
 
References
  • Baumer, E. P. S., Ames, M. G., Burrell, J., Brubaker, J. R., & Dourish, P. (2015). On the importance and implications of studying technology non-use. Interactions, 22(2), 52–56.
  • Bovens, M., & Zouridis, S. (2002). From street-level to system-level bureaucracies: How information and communication technology is transforming administrative discretion and constitutional control. Public Administration Review, 62(2), 174–184.
  • Eubanks, V. (2018). Automating inequality: How high-tech tools profile, police, and punish the poor. St. Martin’s Press.
  • Jasanoff, S., & Kim, S.-H. (2015). Dreamscapes of modernity: Sociotechnical imaginaries and the fabrication of power. University of Chicago Press.
  • Raghunathan, K., et al. (2019). 
  • Raghunathan, K., & Kalyanpur, N. (2021). 
  • Satchell, C., & Dourish, P. (2009). Beyond the user: Use and non-use in HCI. Proceedings of OZCHI.
  • Wyatt, S. (2003). Non-users also matter: The construction of users and non-users of the Internet. In N. Oudshoorn & T. Pinch (Eds.), How users matter (pp. 67–79). MIT Press.