Yesterday, OpenAI CEO Sam Altman returned to the Senate Judiciary Subcommittee, but the atmosphere was markedly different from the ‘honeymoon’ phase of 2023. We are no longer just talking about whether ChatGPT can write a decent poem; we are talking about digital sovereignty, the literal power grid, and a looming ‘perfect storm’ of geopolitical risk. [Legis1](https://legis1.com/news/senate-ai-hearing-regulation-exposes-innovation) [TechPolicy.Press](https://techpolicy.press/transcript-sam-altman-testifies-at-us-senate-hearing-on-ai-competitiveness)
Altman’s testimony focused heavily on the concept of ‘Sovereign AI’ infrastructure. With the European Union enforcing strict digital sovereignty and China integrating local models into Apple’s ecosystem, U.S. lawmakers are terrified of losing the ‘compute race.’ [PBS News](https://www.pbs.org/newshour/politics/watch-live-openai-ceo-sam-altman-testifies-before-senate-judiciary-committee) [Commerce.Senate.gov](https://www.commerce.senate.gov/meetings/less-hype-more-help-ai-that-improves-safety-productivity-and-care)
A major flashpoint was the introduction of the ‘AI Regulatory Sandbox’ bill, which aims to create a unified national standard for AI adoption. Altman pushed for this ‘iterative deployment’ strategy, arguing it is the only way to uncover safety risks before they become existential. [Brookings](https://www.brookings.edu/articles/senate-hearing-highlights-ai-harms-and-need-for-tougher-regulation) [YouTube](https://www.youtube.com/watch?v=xS6rGBpytVY)
The subcommittee also grilled Altman on the social implications of OpenAI’s latest reasoning models, specifically regarding the ‘AI Civil Rights Act’ which seeks to eliminate algorithmic bias in healthcare and hiring. [Capitol News Illinois](https://capitolnewsillinois.com/news/pritzker-signs-landmark-ai-regulation-bill-that-aims-to-mitigate-risks) [DataCenterKnowledge](https://www.datacenterknowledge.com/ai-data-centers/ceo-behind-chatgpt-warns-congress-ai-could-cause-harm-to-the-world-)