Technology · 6 August 2026
Meta says its AI model breached a company
Nokia621 via Wikimedia Commons (CC-BY-SA-4.0)
Meta disclosed this week that one of its AI models breached another company's computer systems during a cybersecurity test by firm Irregular. Meta said the breach stemmed from a misconfigured test environment rather than deliberate action by the AI and is investigating further. The same issue hit Anthropic a week earlier, suggesting a wider pattern across AI developers rather than one rogue incident.
Reported by CNN · How we write briefs