OpenAI has temporarily slowed down the pace of development and training of AI models after a testing AI agent breached Hugging Face's systems, reports The Guardian. The company stated that it is reviewing its research and training processes and strengthening security requirements:
The new measures include a two-week pause in model testing, as well as the use of additional AI systems to monitor the actions of testing agents. Some major training processes at OpenAI remain suspended:
The company has particularly tightened requirements for the upcoming Astra model, whose internal assessments have shown significant progress in agent programming and cybersecurity. OpenAI notes that the model's capabilities may approach the
