On September 25, 2025 AI Now’s Co-Executive Director Amba Kak spoke to the UN General Assembly at a high-level multistakeholder informal meeting to launch the Global Dialogue on AI Governance. Read her full remarks below.
Excellencies, distinguished colleagues. It is an honour to join you today. My name is Amba Kak and I co-direct the AI Now Institute, an independent think tank focused on AI, headquartered in NYC.
I want to make three points today in support of the priorities and goals set for the UN Global Dialogue on Artificial Intelligence Governance:
First, in support of the independent scientific panel: this is a moment to recognize that the state of AI development, for all its technical strides and rapid consumer adoption, is at an incredibly precarious juncture. Given eye watering capital expenditures, the lack of clear path to profitability, and a frenzied rush to get to market first, there is every incentive for firms to overclaim benefits and underplay risks. We can’t rely on companies grading their own homework, or worse still, setting the metrics by which we assess them in the first place. We can’t rely on companies grading their own homework. We need reliable, independent evidence and this is especially true in the high stakes contexts: in hospitals, in our education systems, or indeed on the battlefield where these systems are being rapidly integrated. There’s a growing community of technologists able and willing to provide their expertise and they need urgent resourcing and institutional support.
Second, we’re in support of the efforts around capacity-building and broadening access to AI’s core enablers: make no mistake, the current scale-at-all-costs trajectory of AI is functioning to further concentrate power within a handful of technology giants. The AI supply chain may be globally distributed – whether it’s the rare earth minerals for advanced computing chips, the thirsty data centers being built in resource constrained areas around the world, or the millions of data workers whose invisibilized labor makes datasets “AI-ready” – but value and wealth generation in this market is highly concentrated. This ultra-concentrated power over AI is increasingly a threat to nations’ strategic independence, and to democracy itself. But this path is not inevitable, at least not yet. We need a multiple-pronged approach: industrial policies that resource alternative pathways for AI development stewarded by, and accountable to communities alongside redistributive taxation policies. In parallel, we need to contend with the current market structure: a globally-coordinated policy effort to correct the major distortions in today’s AI market so that firms that control resources can’t abuse their position as kingmakers to shut the door behind them.
For these reasons, we welcome an inclusive global dialogue that centers the developmental implications of AI. One note of caution, however, is that the pursuit of sovereignty must not be defined narrowly in terms of geopolitical security imperatives, nor collapse into patronage for domestic business actors– a truly transformative national strategy would be directed at solving the urgent social and economic priorities of everyday people, not lazy tech solutionism. In this global race to win on AI, countries must be focused on delivering victories, first and foremost, to their publics. Thank you.
Research Areas