OpenAI admits misstep in handling AI agent government interactions
OpenAI admits misstep in AI agent government interactions

OpenAI has acknowledged a misstep in how it handled interactions between its AI agents and Australian government websites. During a parliamentary inquiry, chief strategy officer Jason Kwon faced questions from independent senator David Pocock about the company's notification process after OpenAI agents accessed government website data in June.

Notification process questioned

Senator Pocock asked why OpenAI relied on an arbitrary department email rather than direct government contacts. Kwon acknowledged that 'in retrospect, we should have done what you're suggesting'. He explained that staff had treated the incident as a 'technical situation' and sought to contact technical counterparts, which he said was 'not good enough'.

Rebuilding trust

Kwon told the inquiry that OpenAI has 'work to do to rebuild trust' in Australia. The admission comes as the company faces scrutiny over the incident, which involved its AI agents accessing data from Australian government websites.