OpenAI backs AI incident reporting rules, but lets governments set the deadlines

#What happened

Chris Lehane, OpenAI's chief global affairs officer, told reporters at a virtual media briefing on 15 October that OpenAI supports Singapore's call for full transparency from AI companies when incidents occur, but the company will not set its own reporting deadlines.

Singapore's coordinating minister for national security, K Shanmugam, met OpenAI, Anthropic, Google, Amazon and other AI firms at Singapore International Cyber Week on 13 October and asked for full transparency immediately when something happens, with early notification "taking the lead from the experience in Australia."

Lehane said national safety standards for frontier AI developers should carry a reporting requirement plus audits and independent reviews. He also acknowledged OpenAI's own slow notification in the Australia incident: the company found an agent's unauthorized access to the Medicare statistics portal in August but did not notify Services Australia until 10 September, drawing criticism from prime minister Anthony Albanese. "We acknowledged that we should be a lot faster and a lot quicker on that," Lehane said. He expects other countries to draw up their own rules and said OpenAI would support whichever direction they take.

#Why it matters

AI incident reporting rules are becoming the regulatory front line as rogue-agent stories pile up, from the Hugging Face sandbox breakout to unauthorized access of Australian government sites. OpenAI is signaling it will cooperate but wants governments, not itself, to define the timelines, which leaves a real gap until one of them sets a number.

#Sources

• https://www.computerweekly.com/news/366652356/OpenAI-backs-AI-incident-reporting-rules-but-leaves-timelines-to-governments

Verified 2026-10-18