Prof G MarketsHow To Actually Regulate AI — ft. Alex Bores
25 Sept 2026 · 62 minThe most effective approach to regulating advanced AI will likely involve a combination of clear liability assignment, independent third-party audits with full access and publication rights, and infrastructure fees that incentivize green energy and grid modernization.
Summary
Alex Bores, co-founder of Who Decides, offers a perspective on AI regulation forged from his experience sponsoring New York's RAISE Act. He highlights the political maneuvering by AI companies to influence regulation and proposes practical, actionable steps for policymakers to mitigate the catastrophic risks of AI, emphasizing accountability, transparency, and a democratic approach to technology governance. The conversation is worth attention for its focus on specific regulatory mechanisms rather than abstract discussions about AI safety.
Interesting ideas
AI Accountability: Clarity Over Placement. The critical issue in AI-related incidents is not necessarily whether the user or developer is held responsible, but that a clear, unambiguous party is assigned liability. This clarity is essential to create incentives for safer AI development and deployment, preventing scenarios like the Hugging Face attack where a clear crime occurred but no one was held accountable, leading to a breakdown in trust and repercussions.
Third-Party Audits: Beyond Self-Regulation. Effective AI regulation requires independent third-party audits that go beyond mere self-assessment. These auditors must have employee-level access to data, the ability to publish uncensored reports, and ideally, be chosen and licensed by the government, without direct financial dependence on the audited AI companies, to avoid conflicts of interest seen in other industries.
Data Centers as AI's Democratic Battleground. Local opposition to data centers is emerging as the first significant democratic resistance to AI, channeling broader public anxieties about the technology. Moratoriums on data center construction can provide a crucial pause for regulators to establish rules that mandate green energy, grid upgrades, and community fees, transforming these infrastructure projects from potential liabilities into environmental and economic assets.
Aligning AI Incentives with Public Good. Instead of outright bans, regulation can leverage AI companies' drive for speed by offering expedited permitting and interconnection queue access to those that adopt stringent standards. This incentivizes a 'race to the top' where companies compete to offer the best environmental, labor, and community benefits, aligning their rapid deployment goals with public welfare.
Questions worth thinking about
If clarity on accountability is paramount, what specific legislative language would best establish this without stifling innovation or overly burdening developers?
How can governments rapidly build the internal technical capacity needed to effectively license and oversee independent AI auditors, given the pace of AI advancement?
Beyond data centers, what other points of intersection between AI and local governance might emerge as future 'democratic battlegrounds' for regulation?
Written by AI from the episode's full transcript. It can get things wrong: if something here looks off, tell us at hello@podpings.com.