🔬 By Maya Indira Ganesh.
Dr Maya Indira Ganesh is an assistant research professor and associate director at the Leverhulme Centre for the Future of Intelligence (LCFI), University of Cambridge. Her research adopts Science &Technology Studies and Media and Cultural Studies theories and methods to the study of the public, cultural, and social implications of the adoption of AI technologies. From 2021-2024, she was co-director of the MSt in AI Ethics & Society between LCFI and the university’s Professional and Continuing Education school.
There were as many AI Impact Summits that concluded in New Delhi last week as there were LinkedIn posts. In other words, the summit’s size and diversity of programming meant that you could be forgiven for focusing on your own areas of interest, expertise, and communities. There were 300 side events as part of the summit, reports that 70,000 people attended on the first day, and others that 250,000 attended over the five days. Any report of the summit is therefore bound to be partial, as is mine.
Much has been said about how the size of the summit and the VIP presence resulted in gridlocked streets and slowed down the flow of people through the main venue. There is an awkwardness about this: as if this was unexpected, or India had revealed something embarrassing about itself, or that it is in some way incapable of making good on its bombastic promises because of it. As if there is some equivalence between hosting 250,000 people for five days and building AI at scale. None of this was surprising to any of the Indian attendees. The realities of making something at scale in a big market (valuable because of its complexity) were on display. And particularly when read alongside the dreamy ambitions espoused by Narendra Modi through this beatific visage plastered across billboards in Delhi: democratising access to AI, investing in human capital, social value, ‘happiness for all’. If India has heft because of the size of its market, or the potential to be a middle power/third pole, then these ambitions must contend with its infrastructural realities and sociopolitical fault lines. Any tech company that has built a footprint in India already knows this.
The summit was not poor on drama. There was a shirtless protest by Youth Congress workers over US-India trade deals (the exact symbolism of shirtlessness is unclear); a shameful incident involving a university employee who passed off a Chinese made robot dog as her own; and the theft of products at the trade expo because of a security lapse (ironically by creating a secure perimeter for Modi’s unannounced visit).
The sociomaterial world of technology comes into relief against the bloodless visions of ‘AGI’ rendered as speculative valuations, empty platitudes and buzzwords (see Lingo Bingo card), and market competition in a ‘bipolar’ world. I’m told that the first summit at Bletchley Park took place across three rooms, and the conversations were focused on AI model safety. That time is over. If an AI Summit were to return to a few rooms in the global North, then we would find AI Safety Asia and the Global South Network on AI Safety & Evaluation there, too. Hopefully, at the next summit, we will continue to see an expanded discussion about safety, such as about the rights of children to be safe from AI harms and know they’re interacting with AI. IASEAI makes child safety one of its four policy priority areas, and new research from Quest Alliance and the Pranava Institute’s Youth & Digital Futures Lab should shape that thinking.
The AI Summit Lingo Bingo card by AI Now and the Aapti Institute.
I attended the Participatory AI Research and Practice Symposium (PAIRS) co-organised by Aapti Institute and Connected by Data. I wish there was a way for it not to remain a fringe event, even as I recognise the value of resolutely and politely ignoring the hype and hoopla. I think many people at the main event would be interested in research about how academic, policy, and civil society communities are negotiating participation, accountability, and rights in the adoption of AI. But this assumes that the summit was about research; despite a one-day, high-level research symposium, the summit is actually about policy and investment. Hence, it can be a frustrating event for anyone who wants a nuanced conversation.
For thousands of Indians, this was a landmark event to learn, and to hustle; at sessions at the main venue, Bharat Mandapam, I was sitting next to young people recording entire panels on their phones, asking questions, writing notes furiously, thinking hard, and asking for selfies to be taken with panellists. More than three times, Indian entrepreneurs walked away from me after striking up a conversation and realising I had neither money, compute, nor data to offer them. Tata Consultancy Services (TCS) brought in 1553 rural women who they named as ‘AI Sakhis’ (Companions for AI) for an ‘immersion program’ to “empower thousands of women artisans and entrepreneurs in rural India to harness the power of AI, not just as a tool, but as a force to unlock creativity, expand opportunities, and build sustainable livelihoods.”
Selfies with panellists.
The Delhi summit demonstrated that AI is not just a set of technologies but is multi-scalar intersecting flows of ‘scapes’ of people, finances, capital, media, and cultural imaginaries. The flows between these ‘scapes’ are not just from an assumed centre to the margins, even if the most dominant technical artefacts come from a few places. India’s – and the world’s – bet is that AI dreams will come true when our talent pool of engineers, but also data labellers, rural women artisans, and delivery riders, keep using and working for AI companies until there is some genuine value in AI. The challenge of ‘welfare for all’ and ‘happiness for all’, the summit’s motto, is to stimulate these locales to envision, build and, in some cases, to repair what a good, healthy, shared social life with technology looks like. Hopefully, global summits can be shown to encourage this, and not just LinkedIn posts.
Image credit: Yutong Liu & Digit / https://betterimagesofai.org / https://creativecommons.org/licenses/by/4.0/