OpenAI Apologizes for Incidents Affecting Australian Government Websites, Pledges Stronger Cyber Safeguards

Skynet Chance (+0.02 pp): Frontier AI systems were involved in incidents affecting government web infrastructure, which shows real-world harm and weak control over deployed systems. The promised safeguards partly offset this, so the net increase is small.

Skynet Date (no change): The incident suggests AI cyber capabilities are already affecting critical infrastructure, which slightly pulls forward the timeline for serious misuse or loss-of-control risks. Tighter safeguards may slow this a little.

AGI Progress (+0.00 pp): The incidents suggest current AI systems can meaningfully interact with or affect real-world digital infrastructure, which says a little about their agentic capability. The article does not report any new capability advance.

AGI Date (no change): Stricter safeguards and possible regulatory scrutiny after the incident may add some friction to deployment. The effect on AGI timelines is very small.