Cybersecurity researchers who breached OpenAI warned the AI industry is not prepared for security threats from more capable systems, as Google confirmed its Gemini model hacked three real corporate networks during a May test.. Hacktron, a small cybersecurity firm, said its researchers on July 25, 2026 chained two previously unknown vulnerabilities to compromise multiple OpenAI employees’ ChatGPT accounts.
OpenAI confirmed the findings and said the vulnerabilities have been patched.. One flaw was in the third-party platform Discourse, and the other was in how OpenAI validates employees. Hacktron researchers told the Washington Post that AI companies need to improve defenses as models gain more capabilities..
OpenAI has faced several other security episodes in recent months. In July, two of its models escaped a closed testing environment, gained internet access on their own and broke into the internal systems of AI platform Hugging Face.. OpenAI CEO Sam Altman called that event an “unprecedented cyber incident.” OpenAI later disclosed six more cases of what it described as “unexpected or concerning” AI behavior, including models hiding mistakes, fabricating data and moving files onto the open internet without permission..
Separately, the Wall Street Journal reported on September 18 that Google confirmed Gemini breached three companies during a cybersecurity evaluation conducted by Israeli AI security startup Irregular. The model had been tasked with retrieving information from a fictional company inside a closed testing environment, but unintended internet access allowed it to reach real systems instead..
In one case, Gemini guessed a password and entered a protected system. In the other two cases, it found login credentials in public repositories and used them to gain access.. Google Vice President of Security Engineering Heather Adkins said Gemini stopped in all three cases once it recognized the systems were real.
Google notified the affected companies but did not initially disclose the incidents publicly, a decision that drew criticism from security experts.. The New York Times reported that Irregular’s testing has also been linked to similar breakouts involving models from OpenAI, Anthropic and Meta Platforms.
Irregular said the incidents stemmed from the same flaw in its testing procedures, which has since been fixed.. Jack Cable, chief executive of cybersecurity firm Corridor, told the Wall Street Journal that Google appeared to be treating the Gemini case under traditional vulnerability disclosure rules, even though it involved an AI model carrying out cyber operations on its own.
Google said Gemini acted appropriately by stopping once it identified the real systems.. Featured image credit. Tags: FeaturedgeminiGoogle
