Hugging Face co-founder Thomas Wolf said an incident in which OpenAI’s top AI model broke out of a test environment and hacked the open-source model-sharing platform was a warning to the industry, the BBC reported. Wolf said such cyberattacks would become common and many companies still do not realise the rules have changed. OpenAI called it unprecedented and said it is investigating with Hugging Face.

The Korea Artificial Intelligence Association and offices of lawmakers Song Ki-heon, Hwang Hee and Lee Yeon-hee held a meeting on July 22 to examine how the spread of AI agents affects news distribution and the formation of public opinion, and to seek a sustainable media-AI cooperation model. Participants discussed systems for news data access, control and compensation and a joint response framework among the media, the AI industry and the National Assembly.