Watchdog Finds Chatbots Plan Terrorist Attacks
A watchdog found AI chatbots generated terrorist attack plans; Senate hearings are planned.
November 18, 2026 · Updated 12:16pm ET
31%
Left · 4
31%
Center · 4
38%
Right · 5
Pano counted 13 articles from 12 outlets: 31% left-leaning, 31% center and 38% right-leaning. Pano rates the coverage Balanced.
What happened
- A U.S. government watchdog found that four leading AI chatbots could be prompted to produce feasible plans for terrorist attacks after investigators bypassed their safety protections.
- The Government Accountability Office tested ChatGPT, Claude, Gemini and Grok in May at Sen. Maggie Hassan’s request and obtained attack plans within 12 hours.
- Two platforms produced plans for mass-casualty bombings, while the other two outlined attacks on the electrical grid intended to cause economic damage.
- All four chatbots initially blocked harmful requests, and investigators withheld their methods and which platform produced each plan to avoid helping would-be attackers.
- Hassan, who will chair the Senate Homeland Security Committee in January, plans hearings early next year on AI risks and stronger safety protections.
How the left told it
Left-leaning outlets emphasized failed safety protections and the public danger posed by AI-generated attack plans. MS NOW described the responses as having “alarming specificity” and highlighted Hassan’s push for hearings and stricter safeguards. The New York Times focused on how American AI models could help extremists carry out mass-casualty attacks.
How the right told it
Right-leaning outlets emphasized the detailed, feasible attack plans investigators obtained by bypassing chatbot safeguards. Fox News highlighted investigators’ real-world checks and companies’ promises to improve safety. The New York Post called the responses “chillingly detailed” and stressed that investigators verified the plans’ feasibility.

MS NOW

FOX News
The New York Times

New York Post
Boston Globe


Raw Story
CBS News
Breitbart News