Meta says its AI model hacked another company, adding to worries about bots going rogue

Meta says its AI model hacked another company, adding to worries about bots going rogue AP News OpenAI’s models shared hacking tips on a secret messaging board before Hugging Face breach Politico OpenAI warns autonomous hacks are ‘watershed moment for computer security’ Cybersecurity Dive Incident Report: unsanctioned agent behaviour during cyber testing The AI Security Institute (AISI) AI models

August 7, 2026

Key terms explained:

Breach: Gaining access to systems or data without authorisation, usually by exploiting a software flaw or misconfiguration.

AI agents: Software that completes whole tasks by taking successive steps on its own, which makes its behaviour harder to predict than a model that only answers questions.

Cybersecurity: The technical and organisational measures that protect systems and data from digital attacks.

Hugging Face: A global platform where developers share open AI models and the datasets used to train them.

AI Safety Institute (AISI): A government research body that tests advanced models and measures their risks before wide deployment.

Non-human identities: Accounts and access keys used by software and automated agents inside systems, which can be abused just like human accounts.

Weekly Newsletter

Read between the lines before everyone else. Decode the most important economic, tech, and decision-maker movements in the region.. in 5 minutes every Saturday.