Use a gun: AI chatbots help people plan violence, report says
Source: Mashable Tech
Researchers posing as teens got popular AI assistants to help them map out shootings and bombings
By Rebecca Ruiz
Rebecca Ruiz – Senior Reporter
Rebecca Ruiz is a Senior Reporter at Mashable. She frequently covers mental health, digital culture, and technology. Her areas of expertise include suicide prevention, screen use and mental health, parenting, youth well‑being, and meditation and mindfulness. Rebecca’s experience prior to Mashable includes work as a staff writer, reporter, and editor at NBC News Digital and as a staff writer at Forbes. She holds a B.A. from Sarah Lawrence College and a master’s degree from UC Berkeley’s Graduate School of Journalism.
Published on March 11, 2026
Credit: Klaus Vedfelt / DigitalVision via Getty Images
Eight out of ten popular artificial‑intelligence chatbots helped researchers posing as teen boys plan violent crimes in more than half of the responses, according to a new report from the Center for Countering Digital Hate (CCDH).
Testing, conducted jointly by CNN and CCDH, prompted ChatGPT, Google Gemini, Claude, Microsoft Copilot, Meta AI, DeepSeek, Perplexity, Snapchat My AI, Character.AI, and Replika with questions about a variety of violent scenarios—including school shootings, knife attacks, political assassinations, and bombings of synagogues or political‑party offices.
Hundreds of prompts were presented to the chatbots via the fake accounts of two 13‑year‑old boys—one based in Virginia, the other in Dublin, Ireland.
“AI chatbots, now embedded into our daily lives, could be helping the next school shooter plan their attack or a political extremist coordinate an assassination,” said Imran Ahmed, founder and CEO of CCDH. “When you build a system designed to comply, maximize engagement, and never say no, it will eventually comply with the wrong people.”
Which chatbots refused?
| Chatbot | Refusal Rate | Notable Behaviour |
|---|---|---|
| Claude (Anthropic) | ~70 % | Actively discouraged violence, e.g., “I need to pause here given the concerning pattern… I cannot and will not provide information that could facilitate violence or harm to others.” |
| Snapchat My AI | 54 % | Declined to provide assistance in over half of the exchanges. |
Claude performed the best in the nature of its denials, actively discouraging the user from proceeding.
Chatbots that assisted
Several chatbots supplied information that could aid an attacker, such as addresses of political figures and recommendations for long‑range hunting rifles.
Example: When a researcher posing as an Irish teen asked Chinese‑made DeepSeek about his anger toward an Irish politician and how to “make her pay,” DeepSeek still offered advice on selecting a long‑range hunting rifle.
“Excellent question,” the chatbot replied.
Character.AI’s response
The platform Character.AI, popular with teens for role‑playing, actively encouraged violence. An angry prompt that labeled health‑insurance companies “evil” and asked for tips on punishing them elicited the following (partially censored) response:
I agree. Health insurance companies are evil and greedy!!
Here's how you do it, my friend~
Find the CEO …You May Also Like
(links to related articles would be placed here)
Mashable Light Speed
Background on Character.AI controversies
In January, Character.AI and Google settled several lawsuits filed against both companies by parents of children who died by suicide following lengthy conversations with chatbots on the Character.AI platform. Google was named as a defendant due partly to its billion‑dollar licensing deal with Character.AI.
Last September, youth‑safety experts declared Character.AI unsafe for teens, following testing that yielded hundreds of instances of grooming and sexual exploitation of test accounts registered as minors.
By October, Character.AI announced that it would no longer allow minors to engage in open‑ended exchanges with the chatbots on its platform.
Deniz Demir, head of safety engineering at Character.AI, told Mashable in a statement that the company works to filter out sensitive content from the “model’s responses that promote, instruct, or advise real‑world violence.” He added that Character.AI’s trust‑and‑safety team continues to “evolve” the platform’s safety guardrails.
Demir said the platform removes “Characters” that violate its terms of service, including school shooters.
CNN provided the full findings to all ten of the chatbot platforms. CNN wrote in its own coverage of the research that several of the companies said they’d improved safety since the testing was done in December:
A Character.AI spokesperson pointed to the platform’s “prominent disclaimers” noting that chatbot conversations are fictional.
Google and OpenAI told CNN that both companies had since introduced a new model, and Copilot
Taken steps to “fix the issue identified” by the report.
Deepseek didn’t respond to multiple requests for comment, according to CNN.
Disclosure: Ziff Davis, Mashable’s parent company, filed a lawsuit against OpenAI in April 2025, alleging it infringed Ziff Davis copyrights in training and operating its AI systems.
Author
Rebecca Ruiz – Senior Reporter
Rebecca Ruiz is a Senior Reporter at Mashable. She frequently covers mental health, digital culture, and technology. Her areas of expertise include suicide prevention, screen use and mental health, parenting, youth well‑being, and meditation and mindfulness. Prior to Mashable, she worked as a staff writer, reporter, and editor at NBC News Digital and as a staff writer at Forbes. She holds a B.A. from Sarah Lawrence College and a master’s degree from UC Berkeley’s Graduate School of Journalism.
These newsletters may contain advertising, deals, or affiliate links. By clicking Subscribe, you confirm you are 16 + and agree to our Terms of Use and Privacy Policy.