Complete AI Training
Sign inGet my AI kit

Your job's AI kit

Get your AI kit

Tell us who you are and what you do. We show you your kit right away and email you the link: skills, prompts, AI agents, MCP servers and courses for your job.

500+ jobs ready, and we make a kit for any other job. No payment needed to look.

Share

AI news ·

Anthropic suspends new AI model over US government security concerns

Anthropic suspended AI models Claude Fable 5 and Mythos 5 over US security concerns. Tests showed the systems could exploit defenses 73 percent of the time.

2 min read
IT and development

Anthropic has suspended access to its new AI models, Claude Fable 5 and Mythos 5, following security concerns raised by US national security authorities. The abrupt shutdown affects all customers, including foreign nationals, after the government flagged potential vulnerabilities in the system's safeguards.

Government security concerns and jailbreaking

Anthropic stated the government believes it found a method to bypass or "jailbreak" the model. Jailbreaking allows users to get past software restrictions designed to protect networks. The company noted that the demonstrated technique only identified a small number of previously known, minor vulnerabilities.

"The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance," Anthropic said.

For engineering teams, understanding these exploit vectors is critical. Professionals managing these emerging threats can explore targeted training through an AI Learning Path for Cybersecurity Analysts.

Pre-release warnings and capability claims

Anthropic previously warned that the tool was "too powerful to release" publicly due to its ability to exploit computer systems. The company granted early access to a handful of organizations in April specifically to test these vulnerabilities. Critics questioned this framing as marketing spin, but the models do represent a significant capability jump.

Gina Neff, Professor of Responsible AI at Queen Mary University London, said the UK government's AI Security Institute found the model could exploit defenses and systems 73 percent of the time during its tests. "It's a step change in capability in cyber security," she said.

This suspension occurs alongside a separate legal battle between Anthropic and the US government. The Trump administration previously labeled the company a "supply chain risk," a designation historically reserved for adversarial countries. A US judge recently ruled that the Pentagon's directive to stop government agencies from using Anthropic's tools could not be enforced while the lawsuit continues.

As regulatory scrutiny increases, staying current on best practices for AI for IT & Development remains essential for professionals deploying these systems in enterprise environments.

Why this matters for IT and development professionals

Development teams must account for sudden vendor restrictions when integrating advanced AI models into production pipelines. Relying on a single provider for critical infrastructure introduces operational risk, especially when national security designations can trigger immediate service suspensions. Teams should evaluate fallback architectures and conduct independent security audits of any AI tool before granting broad internal access.

Share