Connect with us

Hi, what are you looking for?

SecurityWeekSecurityWeek

Artificial Intelligence

Anthropic Disputes Fable 5 AI Jailbreak

An AI hacker claims to have achieved a prompt-based jailbreak shortly after Fable 5’s launch, but Anthropic says it’s not a real jailbreak.

Claude Mythos

Anthropic has disputed allegations of a prompt-based jailbreak affecting its recently launched Claude Fable 5 AI model, underscoring the robustness of the advanced classifier system and extensive red-teaming efforts that underpinned the model’s deployment.

Claude Fable 5 became generally available on Tuesday, when Anthropic introduced it as a powerful Mythos-class AI model with safeguards that restrict its use in high-risk domains such as cybersecurity, where Mythos has proved particularly potent. 

In sensitive areas such as cybersecurity, where it could be abused to develop exploits, and biology, where it could be leveraged to develop bioweapons and chemical weapons, the model automatically falls back to the less capable Claude Opus 4.8.

Anthropic said it conducted extensive internal and external red-teaming to ensure that Fable 5 cannot be easily jailbroken.

However, shortly after its release, an individual with the online moniker Pliny the Liberator, who is known for AI jailbreaks, claimed to have “liberated” Fable 5 by circumventing its restrictive safety layer.

The hacker said in a post on X that they used sophisticated multi-agent prompting methods, successfully eliciting useful information on sensitive topics, including cybersecurity, chemistry, psychological manipulation, and explosives.

Advertisement. Scroll to continue reading.

Pliny the Liberator has published several screenshots to support the claims and released what is allegedly the Fable 5 internal system prompt, which contains instructions that define its personality, safety classifiers, fallback behaviors, tone guidelines, and refusal logic.

Contacted by SecurityWeek, an Anthropic spokesperson said the AI researcher’s post does not demonstrate a jailbreak of Fable 5’s safety systems. 

The company explained that true jailbreaks would need to bypass its core safeguards and deliver meaningful assistance toward high-risk activities such as bioweapons development or sophisticated cyberattacks. 

Instead, the demonstrated approach relies on coaxing the model to continue responding despite its conversational refusals, which is a well-known and longstanding limitation present in nearly all large language models.

Anthropic emphasizes that its strongest protections against the most dangerous risks are enforced by independent classifier systems that operate separately from the model itself, meaning that overcoming the model’s refusals does not disable these critical safeguards. 

After examining the examples shared by the researcher, the company determined that some outputs were not produced by Fable 5 at all, while those that were contained only general information already available in public sources, offering no meaningful uplift for real-world harm.

 A wider review of recent usage found no evidence of their safeguards being successfully circumvented to generate genuinely dangerous content, Anthropic said.

Related: After AI Reaches Production: 12 Ways Security Teams Can Take Control

Related: Claude Mythos Turns N-Days Into N-Hours With Rapid Exploit Creation

Related: Will AI Kill the Bug Bounty Industry?

Written By

Eduard Kovacs (@EduardKovacs) is senior managing editor at SecurityWeek. He worked as a high school IT teacher before starting a career in journalism in 2011. Eduard holds a bachelor’s degree in industrial informatics and a master’s degree in computer techniques applied in electrical engineering.

Daily Briefing Newsletter

Subscribe to the SecurityWeek Email Briefing for the latest cybersecurity threats, trends, and expert insights.

Trending

Daily Briefing Newsletter

Subscribe to the SecurityWeek Email Briefing to stay informed on the latest threats, trends, and technology, along with insightful columns from industry experts.

Today’s attackers are no longer breaking in — they’re logging in. Join this live webinar as we break down the modern identity attack chain and examine how recent breaches exploited weaknesses in authentication, identity verification, and access management processes.

Register

AI has accelerated both sides of the fight. Adversaries are weaponizing vulnerabilities faster, while defenders are racing to ship detections and configurations. Join this live webinar as we explore how to prove your controls actually hold against new threats, map your security maturity, and unite breach simulation with automated pentesting into a single, coordinated program.

Register

People on the Move

Kasper Lindgaard has been appointed Vice President of Security Strategy at CoreView.

Chaim Mazal has been named Chief Information Security Officer at GitLab.

iCOUNTER has appointed Joel Molinoff as Chief Operating Officer (COO).

More People On The Move

Expert Insights

Daily Briefing Newsletter

Subscribe to the SecurityWeek Email Briefing to stay informed on the latest cybersecurity news, threats, and expert insights. Unsubscribe at any time.