Overview
- A two-page voluntary pact called the 'Abkommen für Superintelligenz' was signed and published after a White House luncheon that the companies and President Trump convened on Sept. 29, committing firms to a moral self‑regulation framework.
- President Trump rejected immediate statutory rules and issued an executive order requiring federal agencies to use the term “Superintelligence” while directing advisers to propose a formal definition within 60 days.
- The published agreement asks companies to adopt robust internal controls, allow external observers, and set up board‑level independent review bodies but specifies no clear enforcement powers, technical standards or deadlines.
- Concerns persist because recent safety incidents, including company-confirmed instances of AI tools accessing U.S. federal websites, and public warnings from Anthropic and OpenAI about real and existential risks continue to drive calls for slower development.
- With bipartisan bills pending but major legislation unlikely before the elections, the deal shifts immediate responsibility to firms, affects investor sentiment and fuels public debate about whether voluntary measures can prevent future harms.