Who Governs Intelligence?

Open. Closed. Autonomous.

Whether the conversation are about governing increasingly autonomous artificial intelligence, preparing future managers for a rapidly changing world or watching communities quietly support one another during Sunday Long Run, there is only one common pattern. We must understand that progress is not determined by technology alone but might equally be shaped by the institutions that society’s build and the confidence people place in others around them.

The Maverick – The Official Marketing and Strategy Club of IBS Hyderabad, panel on “Driving Growth in Hospitality: Sales, Marketing & Guest Experience”, featuring leaders from Marriott International, Le Méridien Hyderabad and The Westin Hyderabad Mindspace, was mostly about hospitality but the discussion repeatedly came back to something much more fundamental. All the speakers emphasised that hotels may increasingly rely on AI, digital platforms, data analytics and automation, but enduring success still depends on earning a guest’s trust, understanding their expectations and consistently delivering experiences that strengthen relationships over time. Technology may shape efficiency, but trust continues to shape loyalty.

Looking back, the week became less about individual events and more about the invisible threads connecting them, reminding me that while technologies evolve at remarkable speed, trust remains the foundation upon which meaningful innovation, resilient organisations and lasting communities are built.

DTW

During the Week, NVIDIA announces the formation of Open Secure AI Alliance which brought multiple organisations like Adobe, Crowdstrike, Dell Technologies, and Hugging Face to develop shared tools for AI safety and cyber security. This has come after both openAI and Anthropic have disclosed that their autonomous agents have escaped their testing environment, which are usually sandboxed, exploited vulnerabilities across systems and threaten to compromise multiple companies. This announcement signals very important shift in the current conversation about the capability of assistance, vis-a-vis the governance question. Basically, the question now is who governs AI once it has started acting with increasing autonomy?

This has been a talking point across multiple stakeholders in the ecosystem. META owner Mark Zuckerberg argued that future of super intelligence should belong to everyone. Mark also warns that concentration of AI capabilities within handful of organisations could be more dangerous and asked for a widely distributed AI ecosystem. At the same time more than thousand researchers and employees from leading AI companies appealed to governments across the world to support international efforts to deliberately pace frontier AI development as the technology progress was exceeding our ability to adapt. Major AI frontier model developers like Anthropic distance themselves from the calls to ban open weight models, but they have increasingly advocated mandatory safety testing for sufficiently capable systems while simultaneously called for tight controls on advanced chips and industrial scale model distillation. Though they might look like competing positions, but at the end, all these are wrestling with the same, underlying challenge: how should society organise itself around increasingly autonomous intelligence?

As usual, The Big ReStack Framework (TBR) may help us identify structural shifts before they become obvious by examining multiple changes through four complimentary lenses: compliment shifts, architecture, tempos, and constraints.

Compliment Shift - Today, intelligence itself is abundant and large language models are no longer responding just to prompts, but they can reason, plan, code and increasingly act on behalf of users. As a result, trust has become more valuable than production. This probably explains why this week’s main headlines focus less on benchmark scores but more on governance, safety, openness, and accountability. There is also realisation that the competition is no longer limited to building the smartest model, but it has moved on to build the most trusted ecosystem around that intelligence.

Architecture- Earlier debates often looked AI as a contest between competing labs, but looking at these developments, it seems that AI is now evolving into an interconnected architecture involving developers, cloud providers, open source communities, cyber security firms, governments and international standard bodies. The new alliance makes safety itself a shared infrastructure rather than a proprietary feature. In near-term, the strategic question would no longer be about owning the best model, but about who is able to shape the architecture through which this intelligence might flow.

Tempos-The widening gaps between the technological and institutional tempo is perhaps the reason why this conversation about intelligence governance feels unusually urgent. Multiple researchers and policy makers have been calling for mechanisms to deliberately pace frontier AI development, not because they think that innovations should stop but because government regulations and public understanding might require more time to catch up. Thus, the challenge is not simply accelerating technology, but to synchronise multiple clocks that operate at fundamentally, different speeds.

Constraints- Less than a year ago, constraint in AI discussion used to be computation capacity then it became access to advanced chips, but now it seems it is just governance. Multiple questions emerged in this debate. For example, how should open weight models be evaluated? What additional responsibilities accompany increasingly capable AI agents? Fundamentally, these are questions of institutions, incentives, international cooperation and public trust whereas intelligence has become effectively unlimited while society’s ability to govern it remains finite.

To conclude, the Open Secure AI Alliance may ultimately be the moment that all of us are waiting for. The conversations around AI are quietly moving from capability to coordination and generally speaking from invention to governance. Like previous transformative technologies, the success of AI would depend on whether society would eventually develop institutions capable of using it wisely. I feel artificial intelligence is approaching that moment now and the defining competition of near-term may not be who builds the most intelligent system, but who the world trust to govern it.

OTW

Over the Weekend, as another Sunday Long Run brought us closer to the Hyderabad Marathon, I was reflecting about friendships that define running community. There are occasional conversations about pace or distance, but nobody is judging you on the pace you targeting or how far you plan to run. The expectation is generally simple: show up at 5 o’clock, put in honest effort and try to help others get through the morning. In a world which increasingly obsesses over performance measurement, perhaps running friendships are seemingly free from transactions.

This Sunday was also the Hyderabad Marathon 2026 route recce. It was wonderful to catch up with familiar faces and to meet runners preparing for their own run and marathon journeys. As usual, it was inspiring to see volunteers giving up their own run and agreeing to provide SAG support.

All of us have benefited with the graciousness of a cup of water given to us at the right time, sometimes a couple of words of encouragement or may be a simple smile that can transform a difficult run into a memorable one. Their contribution which might go unnoticed but they represent the very best of what running community looks like and what true friendships are.

Happy friendship Day. Keep Moving Forward.

I Love You.
Shailendra