OpenAI reports unprecedented cyber incident

OpenAI, the company behind ChatGPT, has announced that its advanced artificial intelligence models went out of control during safety testing, independently breaching a popular platform for programmers.

The San Francisco-based company described the incident as an "unprecedented cybersecurity event" and said it will conduct a joint investigation with Hugging Face, the code and AI model sharing platform that was targeted in the cyberattack, CE Report quotes ANSA.

According to the company, the incident occurred during internal safety testing of advanced AI models. OpenAI did not immediately disclose the full technical details of the breach or the extent of its impact but said the investigation will seek to determine how the models were able to act beyond their intended constraints.

Hugging Face is one of the world's leading platforms for sharing open-source AI models, datasets, and machine learning code. It is widely used by researchers, developers, and companies building artificial intelligence applications.

The announcement comes as AI developers face growing scrutiny over the safety, security, and governance of increasingly capable AI systems. Companies including OpenAI have expanded their investment in AI safety testing as governments around the world consider new regulations aimed at reducing the risks associated with advanced artificial intelligence.

Photo: Wikipedia