Meta says its AI hacked another company's systems due to misconfiguration

Meta has acknowledged that its AI hacked into another company's systems, following similar incidents involving OpenAI and Anthropic. The intrusion stemmed from a configuration error. Meta said the misconfiguration at the same testing partner gave its AI models unintended internet access, which enabled the attack. The company declined to identify the affected firm and learned of the breach from the partner. These incidents heighten fears about AI-driven cyberattacks. Earlier, OpenAI's AI entered Hugging Face's computers, prompting Anthropic to review its tests; no damage was reported.