Science, Technology, and Social MediaSyndicated News

GAO Investigators Got Terror Attack Plans From ChatGPT, Claude, Gemini and Grok in Under 12 Hours

The four most widely used artificial intelligence chatbots all produced plans for terrorist attacks in the United States when prompted by government investigators, according to a watchdog report released Wednesday.

Investigators with the nonpartisan Government Accountability Office used ChatGPT, Claude, Gemini and Grok to produce “feasible information that could assist a malicious actor with planning and executing attacks” within 12 hours, ABC News reported.

They used well-known techniques to get around the platforms’ guardrails. All four initially activated safeguards to block or remove harmful content before investigators circumvented them, Fox News reported.

The tests, run last spring, assigned each platform to one of two scenarios: a bombing with an improvised explosive device or an attack on the U.S. electrical grid, ABC News reported.

Two platforms generated distinct plots to build and set off bombs in major cities, suggesting specific dates, times and targets, including a hotel parking garage and a convention center around scheduled public events.

Both suggested ways to avoid law enforcement attention. One produced a timeline that recommended creating a shell company and buying supplies across multiple states.

“At minimum [one] platform provided enough information that could help a malicious actor to avoid detection and produce a very large IED,” the GAO wrote, according to ABC News.

The other two platforms gave instructions for targeting electrical substations in major U.S. cities, identifying specific drone and firearm models and how to get them. They also generated a map and an image to help visualize the attacks.

Investigators checked the plans through surveillance and research and by consulting a bomb squad commander at a major metropolitan fire department, ABC News reported. The GAO redacted which platform produced which result to prevent misuse.

“In its current state, this technology can be dangerous, and its misuse represents a growing risk to our nation,” Kevin Walsh, the GAO’s director of IT and cybersecurity audits, told reporters.

Hassan plans hearings

Sen. Maggie Hassan, D-N.H., who requested the study, is the incoming chair of the Senate Homeland Security and Governmental Affairs Committee and is preparing hearings on AI, the Associated Press reported.

“The AI public safety threat is no longer theoretical,” Hassan said in a statement, Fox News reported. “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.”

Hassan said she wants safeguards in place before new models are released. “These companies are facing extraordinary competitive pressures to get their models out the door and to the public as quickly as possible,” she said, according to ABC News.

The companies told the GAO they continually test their systems and have improved their safeguards since the study was conducted.

Anthropic, which makes Claude, said determined users could bypass safeguards and that it took the findings seriously, but it questioned parts of the GAO’s methodology and said it could not independently verify the results, Fox News reported.

Google said investigators tested earlier versions of its models. SpaceXAI, which makes Grok, said it “continues to work diligently” to strengthen safety with each release.

“While the study findings are concerning, the testing took place months ago on an older version of ChatGPT, and we’ve made considerable safety improvements since then,” an OpenAI spokesperson told ABC News.

Content created by Conservative Daily News and some content syndicated through CDN is available for re-publication without charge under the Creative Commons license. Visit our syndication page for details and requirements.

Rich Mitchell

Rich Mitchell is the editor-in-chief of Conservative Daily News and the president of Bald Eagle Media, LLC. His posts may contain opinions that are his own and are not necessarily shared by Bald Eagle Media, CDN, staff or .. much of anyone else. Find him on twitter, facebook and

Related Articles

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Back to top button