A few weeks back i watched a documentary named “The thinking game” that dived deep into how Demis Hassabis and his team at Google DeepMind are training the next generation of AI, how he stumbled upon the idea of AGI (Artificial General Intelligence), their innovation in protein folding (one of the most rewarding discoveries by AI) and how they’re developing environments to create sentient AGI.
After this I stumbled upon another documentary called I’m not a robot — a 2023 short by Dutch director Victoria Warmerdam — in which a woman named Lara sits at her desk and tries to tick a simple checkbox: “I’m not a robot.” She fails. She tries again. She fails again. Tech support tells her, gently, that she wouldn’t be the first to find out this way. By the end, a system has concluded there is an 87% chance she is a bot. Her memories, her relationships, her sense of self — all placed in question by an algorithm that never asked her what she thought.
That sent me into an anxious spiral of interviews and explainer videos, like:
- Day after AGI at Davos- Where the top minds in AI talk about the proposed timelines to reach AGI, and the range goes from the end of 2026 to a few years out. Amodei told the audience that AI models would replace the work of all software developers within a year and reach Nobel-level scientific research in multiple fields within two years. Hassabis, more cautious, put the timeline for genuine human-level AGI at five to ten years — but acknowledged a 50% chance it arrives within the decade.
- AI Slop– On how AI is trained on existing data and if going ahead we train more models on data that is now being generated by people using AI — where does that lead us? And the hallucination and tendency of AI to create links that are fake or validate even the most random claims to provide an answer. (There was a joke earlier that you could use Google to prove anything by searching for reports proving your POV— e.g., there are papers stating coffee is good for health and then there are ones claiming it is bad— now multiply this with AI and the content it’s creating that goes beyond just text and into the realm of images and videos— what is even true and how do you establish that going forward?)
- Humanity’s final invention– Talking about the day AGI gets created and what is in store for us after that and much more.
You get the larger idea — the race to AGI is an anxious bubble waiting to be burst into reality, sooner than later. And the resources, time and effort being put into it are enormous with a promise to solve everything from world hunger to cancer and beyond — all within our lifetimes.
It captures something that feels increasingly true about the moment we’re in: technology is moving faster than our ability to understand what it means for us as humans — and far faster than any conversation about what it means for the communities we serve.
That’s where it got me thinking about another crucial timeline full of anxiety and unexpected outcomes – the Covid-19 pandemic. The time everything changed and those who changed with it survived while others felt its colossal impact bite deep into years of work.
The impact on NGOs in the Global South was severe and largely invisible to those making the decisions. Field programs shut down overnight. Community trust — built over years of relationship — was disrupted by mandates that didn’t account for how people actually live. Digital health tools rolled were out without considering that the communities they were meant to serve were often offline, multilingual, or living in contexts where “contact tracing” had very different implications.
The wave hit. Many NGOs were not riding it — they were caught under it.
AI has the shape of a similar wave. It is being built at speed, with enormous capital, by a small number of institutions in a small number of countries. Both Amodei and Hassabis warned that risks are rising as systems become more autonomous — pointing to concerns about labour displacement, misuse, and the need for global coordination on safety standards. But “global” in these conversations tends to mean governments and large enterprises. It does not tend to mean an NGO in Jharkhand deciding whether to use an AI chatbot to deliver health information to adolescent girls.
The communities that stand to be most affected by AI — those already navigating poverty, exclusion, and systemic disadvantage — are almost entirely absent from the rooms where these decisions are being made.
But here is what is different this time.
COVID arrived without warning. AI is arriving in real time.
We can see it coming. We have time — not infinite, but real — to decide how we show up.
The NGOs that moved early during COVID — that adapted quickly, maintained community trust, and held relationships when everything else fractured — were the ones that survived and led. The same will be true for AI.
The question isn’t whether AI will reshape how NGOs work. It will. The question is whether we shape that transition ourselves, or have it handed to us by vendors, funders, and policy frameworks that weren’t designed with our communities in mind.
Careful governance and collaboration can steer AI toward broad societal benefit. But that governance won’t happen without the social sector at the table — pushing for open-source infrastructure, community data rights, and AI tools accountable to the people they’re meant to serve.
Lara, in her film, eventually accepts the label the algorithm gave her. She didn’t fight it. She didn’t have the tools or the community to push back.
We do.
—
Back to the learnings from the COVID era and they can help with AI
While the world was in an unpredictable lockdown, I was also locked into an apartment in a new city, having to pivot in my work from physical setups to going digital — for what i was hired to do was to delve deeper into the knowledge within the social sector by having conversations with some of the most intriguing people in the space.
So off I went interviewing many of these brilliant people and understanding what moved them, what shaped their work, how they began, who they began for, how they had persevered so far, and what keeps them going and so forth.
For instance — sitting in her office in Pune, with a fan whirring in the background on a summer day, while recording a podcast with me over a phone call- Prema Gopalan from Swayam Shikhan Prayog narrated the story of how she was able to gather the people around the idea of shram daan or giving by doing (roughly translated) to rebuild an earthquake stricken area with the people themselves spearheading everything from governance to finance to health and beyond — all bound by the idea of building a future they wanted to live in. And as her legacy lives on, so does her work and the waves her organisation has created.
And hers is one of the many stories that had me amazed and spellbound at the sheer dedication, ingenious ideation and resourcefulness of these leaders. From Breakthrough Trust using songs for behavior change, to Katha and Geeta Dharmarajan putting education into story books, to S.D. Shibulal venturing into using education as a leveler for change through Shikshalokam after having led Infosys, to the case study where an NGO by merely providing cycles to girls was able to increase literacy , the stories from the sector are overflowing and so is the knowledge and learning we can draw from each.
—
Let’s Build SGI
Sounds gimmicky for sure but like every new word that trends to garner interest from the right audience by repackaging the same thing in a new format — here is yet another one — Social General Intelligence.
Maybe a platform, maybe a knowledge hub or maybe just a bot created for the social sector built on context from all these real life stories and deeper insights from the lived experience of those who have touched grass and helped grow it over time — rather than relying on assumptions and biases of those who tried to crack the quickest route to learning or the easiest way to build a large community. Because real change does take time and at times a lifetime — the question now is how do we leverage all these collective lifetimes’ worth of experience to build a future worth living in rather than create yet more AI slop.
