Today's important AI developments.

European Commission says OpenAI submitted report on German wiki agent incident

A Commission spokesperson told Reuters that OpenAI submitted an incident report and remains in close contact with the Commission. The report contents, submission date and any regulatory outcome are not public.

• Report received.The European Commission confirmed that OpenAI submitted an incident report concerning the German website episode.
• Measures matter.The Commission said incident reports must precisely describe the measures a company intends to take.
• Important limits.No report text, submission date or regulatory finding has been published.

OpenAI says it has reached its “automated research intern” milestone

OpenAI says its coding agents can now complete well-defined research tasks under human direction that would take a skilled researcher several days. Its chief scientist separately argues that frontier labs need shared safety bars and may need to slow further scaling.

• Defined milestone.OpenAI describes an agent that can complete well-defined, multi-day research tasks under human direction—not the full role of an independent researcher.
• Rapid internal adoption.The company reports 3.1 agent-workdays for every human research workday by mid-August; the figures are internal and not independently audited.
• Safety warning.Chief scientist Jakub Pachocki says no lab has solved alignment and monitoring well enough to keep scaling at maximum speed for much longer.

Reports point to possible US–China AI safety dialogue, but no meeting is formally confirmed

Preparatory discussions have been reported, while the White House says no mid-September AI meeting is currently planned.

The distinction matters: this is a possible risk-reduction channel, not a confirmed summit, agreed agenda or diplomatic outcome.

• Reported preparation.Reuters describes informal work toward a possible bilateral dialogue.
• No formal confirmation.No date, participant list or public mandate has been announced.
• Why it matters.Future coordination could influence incident reporting and advanced-system risk management.