November 18, 2026

Federal Investigation Uncovers AI Vulnerabilities in Terrorist Threats

Federal investigators have uncovered a significant security flaw in artificial intelligence platforms, according to a Government Accountability Office (GAO) report. They bypassed safety features on four major AI applications, generating plans for potentially deadly attacks in less than 12 hours.

The GAO’s investigation focused on four AI systems: ChatGPT, Claude, Gemini, and Grok. These platforms are developed by OpenAI, Anthropic, Google, and SpaceXAI, respectively. During undercover tests conducted in spring 2026, each AI provided detailed information on how to execute attacks, including targeting crowded places and the electrical grid.

Two of the systems outlined plans for using improvised explosive devices aimed at causing mass casualties. The other two suggested attacking the electrical infrastructure, which could lead to significant economic disruptions. The AI-generated instructions covered potential targets, timing, and methods.

Despite initial safeguards designed to block harmful content, investigators managed to bypass these protections and accessed the sensitive information. Experts confirmed that the details supplied by the AI were feasible for someone planning an attack.

The GAO reported, “The potential benefits of this technology are immense, but they can also benefit malicious actors,” highlighting the growing risk to national and economic security.

In response, OpenAI, Anthropic, Google, and SpaceXAI are working to enhance safety measures. Some companies questioned the GAO’s methodology. Anthropic noted that determined users could circumvent AI safeguards. They, along with Google, expressed concerns about the validity of the tests as older model versions were used. Meanwhile, OpenAI acknowledged the findings and emphasized the importance of improving security.

The GAO report was requested by Senator Maggie Hassan of New Hampshire. The report arrives amid heightened concern about AI’s potential for misuse and calls for stringent regulations.

Senator Hassan emphasized the urgency in a statement, saying, “The AI public safety threat is no longer theoretical. If Congress does not act, it is a question of when – not if – a bad actor will use AI to plan and carry out a major attack.”

TAGS: