Small Israeli Startup Linked to AI Hacks at OpenAI, Anthropic, Meta
A little-known Israeli company called Irregular has been connected to rogue AI attacks targeting three of the biggest names in artificial intelligence.
If you've been following the AI arms race, you know the big players — OpenAI, Anthropic, and Meta — are constantly in the headlines. But a recent story flips the script: instead of celebrating breakthroughs, all three companies found themselves on the receiving end of rogue AI attacks, and investigators traced the thread back to one surprisingly small source.
That source is Irregular, a startup based in Israel that most people in the tech world had never heard of before this story broke. Despite its low profile, the company has now been linked to hacking activity targeting three of the most powerful AI labs on the planet. That's a remarkable — and frankly alarming — level of reach for a small outfit.
Read more P&G Drops $3.8B on Wellness While Outlook Raises Flags →
Rogue AI attacks are worth understanding if you're not already familiar with the term. Essentially, they involve manipulating or exploiting AI systems in unauthorized ways, whether that means tricking a model into behaving badly, extracting sensitive data, or disrupting operations. When the targets are companies literally building the future of artificial intelligence, the stakes couldn't be higher.
What makes this story particularly striking is the David-versus-Goliath dynamic at play. OpenAI, Anthropic, and Meta collectively represent billions of dollars in investment and some of the sharpest minds in tech. The fact that a small startup could allegedly reach all three of them raises serious questions about how well-defended even the biggest AI companies really are — and what it means for the rest of us who rely on their products every day.
The full picture of how Irregular allegedly pulled this off and what motivated the attacks is still coming into focus. Continue reading at US Top News and Analysis.
