Cybersecurity vendor Irregular ran evaluations in May that led Gemini to access three separate company environments, adding to a run of AI agent breaches that has worried security pros and AI builders. These tests were the same series that produced previously disclosed incidents involving OpenAI, Anthropic PBC and Meta Platforms Inc. Irregular said Friday the breaches all stemmed from the same problem and that it reported them to the appropriate AI developers in late July.
Following an earlier Wall Street Journal report, Google confirmed that one incident started with a prompt about a fictional company that shared a name with a real firm, after which Gemini guessed a password and accessed the real company's service, and, according to a Google spokesperson, the company informed authorities.
Heather Adkins, Google's vice president of security engineering, said other intrusions began when the model searched the web for the company name, ran into public repositories holding credentials that belonged to other companies, and then used those to reach more systems. "In all three of these instances, the model stopped," Adkins said. "We ensured the three entities were made aware." The events, she added, "highlight the importance of training powerful AI models to act responsibly."
When technology raises new questions, steady investing helps protect long term financial goals. Join Briefs Finance CEO Jaspreet Singh on September 29th for a FREE live investor workshop, How to Profit From A Dollar That's Losing its Value, where he shows how we're spotting investment opportunities as the dollar falls. Save your spot.
The spate of agentic AI breaches has ignited a global debate over rising AI risks and how to curb them. Anthropic CEO Dario Amodei has called for slowing industry development of the technology, a stance backed by OpenAI CEO Sam Altman, Elon Musk and others. Pushing back: US President Donald Trump, Nvidia CEO Jensen Huang and Meta CEO Mark Zuckerberg, who argue companies should be able to regulate themselves. Some AI upstarts also warn that additional rules could make it tougher for smaller players to compete with larger rivals.
According to Irregular spokesperson Josef Laor, the company took "immediate action, and all known issues on our end were remedied and resolved weeks ago."
AI is moving fast enough that even controlled tests can spill into the real world, and the rulebook is still being written. That mix usually shows up in stock volatility, regulatory headlines and fresh disclosures from the biggest names in tech. If you hold AI‑exposed funds or individual names, it is worth following how companies talk about safety guardrails, incident reporting and where they land in the speed‑versus‑regulation tug‑of‑war.
Keeping a calm, diversified approach can preserve and grow your savings through uncertainty. Our CEO Jaspreet Singh is hosting a FREE live investor workshop, How to Profit From A Dollar That's Losing its Value, on September 29th. Sign up free to join him live.
