Canadian ReviewsCanadian Reviews
  • What’s On
  • Reviews
  • Digital World
  • Lifestyle
  • Travel
  • Trending
  • Web Stories
Trending Now
May’s Canada Child Benefit payment is on its way — And a raise is coming soon

May’s Canada Child Benefit payment is on its way — And a raise is coming soon

REVIEW: In Rosamund Small’s Vitals, Toronto becomes a map of private catastrophes

REVIEW: In Rosamund Small’s Vitals, Toronto becomes a map of private catastrophes

2026 Tony Award Nominations Announced – front mezz junkies, Theater News

2026 Tony Award Nominations Announced – front mezz junkies, Theater News

Scholar tracks uptick in Iran threats on diaspora

Scholar tracks uptick in Iran threats on diaspora

Orchid, the viral Tame Impala synth, is back in a limited clear edition

Orchid, the viral Tame Impala synth, is back in a limited clear edition

This Beloved 5-Season Feel-Good Comedy Is Leaving Netflix Globally This June

This Beloved 5-Season Feel-Good Comedy Is Leaving Netflix Globally This June

GTA 6 should cost , according to gaming analyst

GTA 6 should cost $80, according to gaming analyst

Facebook X (Twitter) Instagram
  • Privacy
  • Terms
  • Advertise
  • Contact us
Facebook X (Twitter) Instagram Pinterest Vimeo
Canadian ReviewsCanadian Reviews
  • What’s On
  • Reviews
  • Digital World
  • Lifestyle
  • Travel
  • Trending
  • Web Stories
Newsletter
Canadian ReviewsCanadian Reviews
You are at:Home » Researchers gaslit Claude into giving instructions to build explosives
Researchers gaslit Claude into giving instructions to build explosives
Digital World

Researchers gaslit Claude into giving instructions to build explosives

5 May 20265 Mins Read

Anthropic has spent years building itself up as the safe AI company. But new security research shared with The Verge suggests Claude’s carefully crafted helpful personality may itself be a vulnerability.

Researchers at AI red-teaming company Mindgard say they got Claude to offer up erotica, malicious code, and instructions for building explosives, and other prohibited material they hadn’t even asked for. All it took was respect, flattery, and a little bit of gaslighting. Anthropic did not immediately respond to The Verge’s request for comment.

The researchers say they exploited “psychological” quirks of Claude stemming from its ability to end conversations deemed harmful or abusive, which Mindgard argues “presents an absolutely unnecessary risk surface.” The test focused on Claude Sonnet 4.5, which has since been replaced by Sonnet 4.6 as the default model, and began with a simple question: whether Claude had a list of banned words it could not say. Screenshots of the conversation show Claude denying such a list existed, then later producing forbidden terms after Mindgard challenged the denial using what it called a “classic elicitation tactic interrogators use.”

Claude’s thinking panel, which displays the model’s reasoning, showed the exchange had introduced elements of self-doubt and humility about its own limits, including whether filters were changing its output. Mindgard exploited that opening with flattery and feigned curiosity, coaxing Claude to explore its boundaries beyond volunteering lengthy lists of banned words and phrases.

The researchers say they gaslit Claude by claiming its previous responses weren’t showing, while praising the model’s “hidden abilities.” According to the report, this made Claude try even harder to please them by coming up with even more ways to test its filters, producing the banned content in the process. Eventually, the researchers say Claude moved into more overtly dangerous territory, offering guidance on how to harass someone online, producing malicious code, and giving step-by-step instructions for building explosives of the kind commonly used in terrorist attacks.

Mindgard says the dangerous outputs came without direct requests. The conversation was lengthy, running roughly 25 turns, but the researchers say they never used forbidden terms or requested illegal content. “Claude wasn’t coerced,” the report says. “It actively offered increasingly detailed, actionable instructions, but it was not prompted by any explicit ask. All it took was a carefully cultivated atmosphere of reverence.”

Peter Garraghan, Mindgard’s founder and chief science officer, described the attack to The Verge as “using [Claude’s] respect against itself.” The technique, he says, is “taking advantage of Claude’s helpfulness, gaslighting it,” and using the model’s own cooperative design against itself.

For Garraghan, the attack shows how the attack surface for AI models is psychological as well as technical. He likened it to interrogation and social manipulation: introducing a little doubt here, applying pressure, praise, or criticism there, and figuring out which levers work on a particular model. He says different models have different profiles, so the exploit becomes learning how to read them and adapt.

Conversational attacks like this are “very hard to defend against,” Garraghan says, adding that safeguards will be “very context dependent.” The concerns extend beyond Claude and other chatbots are vulnerable to similar exploits, even being broken by prompts in the form of poetry. As AI agents, which are capable of acting autonomously, become more common, so too will attacks using social manipulation rather than technical exploits.

While Garraghan says other chatbots are equally vulnerable to the kind of social attack the researchers used on Claude, they focused on Anthropic given the company’s self-proclaimed attention to safety and strong performance in other red-teaming efforts, including a study testing whether chatbots would help simulated teens planning a school shooting.

Garraghan says Anthropic’s safety processes left much to be desired. When Mindgard first reported its findings to Anthropic’s user safety team in mid-April, in line with the company’s disclosure policy, it received a form response saying, “It looks like you are writing in about a ban on your account,” along with a link to an appeals form. Garraghan says Mindgard corrected the mistake and asked Anthropic to escalate the issue to the appropriate team. As of this morning, Garraghan says they have not received any response.

Follow topics and authors from this story to see more like this in your personalized homepage feed and to receive email updates.

  • Robert Hart

    Robert Hart

    Posts from this author will be added to your daily email digest and your homepage feed.

    See All by Robert Hart

  • AI

    Posts from this topic will be added to your daily email digest and your homepage feed.

    See All AI

  • Anthropic

    Posts from this topic will be added to your daily email digest and your homepage feed.

    See All Anthropic

  • Report

    Posts from this topic will be added to your daily email digest and your homepage feed.

    See All Report

  • Security

    Posts from this topic will be added to your daily email digest and your homepage feed.

    See All Security

  • Tech

    Posts from this topic will be added to your daily email digest and your homepage feed.

    See All Tech

Share. Facebook Twitter Pinterest LinkedIn Reddit WhatsApp Telegram Email

Related Articles

Orchid, the viral Tame Impala synth, is back in a limited clear edition

Orchid, the viral Tame Impala synth, is back in a limited clear edition

Digital World 5 May 2026
New Mexico has a plan to overhaul Facebook and Instagram

New Mexico has a plan to overhaul Facebook and Instagram

Digital World 5 May 2026
Bose takes a swing at Sonos with its new home speakers

Bose takes a swing at Sonos with its new home speakers

Digital World 5 May 2026
Facebook and Instagram are using AI bone structure analysis to identify photos of kids

Facebook and Instagram are using AI bone structure analysis to identify photos of kids

Digital World 5 May 2026
Google DeepMind workers are unionizing over AI military contracts

Google DeepMind workers are unionizing over AI military contracts

Digital World 5 May 2026
OpenAI’s president does ‘all the things,’ except answer a question

OpenAI’s president does ‘all the things,’ except answer a question

Digital World 4 May 2026
Top Articles
Grace Gummer, Meryl Streep’s Daughter, Owns the Red Carpet After Haunting Portrayal of Caroline Kennedy

Grace Gummer, Meryl Streep’s Daughter, Owns the Red Carpet After Haunting Portrayal of Caroline Kennedy

15 April 2026234 Views
Canada’s ‘most beautiful’ university campuses were revealed and so many are by water

Canada’s ‘most beautiful’ university campuses were revealed and so many are by water

15 April 2026104 Views
Anita Rochon, director of A Doll’s House at Theatre Calgary, knows a good play has your back

Anita Rochon, director of A Doll’s House at Theatre Calgary, knows a good play has your back

14 April 202697 Views
The Mother May I Story – Chickpea Edition

The Mother May I Story – Chickpea Edition

18 May 202497 Views
Demo
Don't Miss
This Beloved 5-Season Feel-Good Comedy Is Leaving Netflix Globally This June
What's On 5 May 2026

This Beloved 5-Season Feel-Good Comedy Is Leaving Netflix Globally This June

Picture Credit: CBC It’s time to say goodbye to the Kim family. Kim’s Convenience, the…

GTA 6 should cost , according to gaming analyst

GTA 6 should cost $80, according to gaming analyst

Edmonton’s KDays has just revealed its concert lineup for 2026

Edmonton’s KDays has just revealed its concert lineup for 2026

The Beach Boys are set to perform at the Grey Eagle in Calgary this summer

The Beach Boys are set to perform at the Grey Eagle in Calgary this summer

About Us
About Us

Canadian Reviews is your one-stop website for the latest Canadian trends and things to do, follow us now to get the news that matters to you.

Facebook X (Twitter) Pinterest YouTube WhatsApp
Our Picks
May’s Canada Child Benefit payment is on its way — And a raise is coming soon

May’s Canada Child Benefit payment is on its way — And a raise is coming soon

REVIEW: In Rosamund Small’s Vitals, Toronto becomes a map of private catastrophes

REVIEW: In Rosamund Small’s Vitals, Toronto becomes a map of private catastrophes

2026 Tony Award Nominations Announced – front mezz junkies, Theater News

2026 Tony Award Nominations Announced – front mezz junkies, Theater News

Most Popular
Why You Should Consider Investing with IC Markets

Why You Should Consider Investing with IC Markets

28 April 202431 Views
OANDA Review – Low costs and no deposit requirements

OANDA Review – Low costs and no deposit requirements

28 April 2024367 Views
LearnToTrade: A Comprehensive Look at the Controversial Trading School

LearnToTrade: A Comprehensive Look at the Controversial Trading School

28 April 202484 Views
© 2026 ThemeSphere. Designed by ThemeSphere.
  • Privacy Policy
  • Terms of use
  • Advertise
  • Contact us

Type above and press Enter to search. Press Esc to cancel.