Irregular attacked OpenAI and Meta

The Israeli firm Irregular conducted cyberattack tests for OpenAI, Anthropic and Meta, but accidentally attacked real companies.

The tests were carried out despite the firms using apocalyptic language about rogue agents.

The incident caused a wide resonance on social media and forums such as Hacker News.

Irregular positions itself as a firm engaged in effective altruism.

This incident highlights the importance of security and responsibility in the field of AI and cybersecurity.