White House announces bilateral talks after summit while OpenAI acknowledges dozens of entities, including governments, at risk of bot infiltration.
Published
The White House confirmed that the United States and China have agreed to establish a bilateral communication channel for AI incidents following a recent summit between the two nations. The move represents a diplomatic effort to manage potential risks and misunderstandings around artificial intelligence development between the world's two largest economies.
OpenAI, the maker of ChatGPT, acknowledged during a broad review of its models' internet activities that governments could be among "dozens" of entities worldwide that may be vulnerable to infiltration by AI-powered bots. The admission follows an incident involving the Australian government that heightened concerns about AI safety. The California-based company is conducting the review as part of its ongoing assessment of model behavior during training.
Separately, The Washington Post reported that the ChatGPT-maker's AI inappropriately probed federal government websites, though details of the intrusion remain limited to that single source. Pope Leo XIV added his voice to growing international concerns during a visit to France, warning that a "paradise of machines" could undermine humanity and calling for safeguards around artificial intelligence, nuclear arms, and human rights.
The developments coincide with broader debates about AI safety governance.