Overview
- A group of roughly two dozen tech leaders met with President Trump and signed a two‑page voluntary agreement on Sept. 29 that commits companies to internal controls, external reviewers, and new independent board committees to oversee AI safety.
- The pact is non‑statutory and described by Mr. Trump as “morally binding,” and he followed the meeting with an executive order instructing agencies to use the term “Super Intelligence.”
- The agreement leaves key enforcement details unspecified, prompting critics and safety advocates to question how the commitments will be monitored or enforced by regulators or courts.
- Some company leaders including Anthropic and parts of OpenAI have urged slowing the most powerful models, but the new White House approach does not impose development pauses or formal limits.
- OpenAI’s Sam Altman was at the company’s DevDay unveiling its Dots agents while Greg Brockman signed for OpenAI, and analysts say Congress is unlikely to pass federal AI laws before the November elections so states and markets will shape near‑term rules and risks.