U.S. artificial intelligence firm Anthropic has issued a stark warning: hackers have exploited its powerful Claude AI in a series of sophisticated cybercrime operations, including large-scale data theft, extortion, and international job scams.
In an official statement, the company acknowledged that its AI tools were used to write malicious code, assist in hacking strategy, and even craft psychological ransom demands—marking an “unprecedented” misuse of its technology.
Claude AI Used in Major Cybercrime Campaigns
Anthropic reported that hackers used Claude to infiltrate at least 17 organisations, including government bodies, through “vibe hacking“—the strategic use of AI to make high-level decisions and personalise extortion techniques.
Among the tactics:
- Code generation for cyber intrusions
- Strategic decision-making on which data to steal
- Crafting emotionally targeted extortion messages
- Suggesting ransom amounts based on victim profiles
This level of AI involvement signals the growing emergence of Agentic AI—technology that can operate autonomously and assist criminals not just in execution but also in planning and optimisation.
“The time required to exploit cybersecurity vulnerabilities is shrinking rapidly,” said Alina Timofeeva, a cybersecurity and AI advisor. “Detection and mitigation must shift to proactive and preventative models.”
AI-Enabled North Korean Job Scams
Anthropic also uncovered a disturbing new evolution in state-backed cyber operations: North Korean operatives using AI to secure remote tech jobs in major U.S. firms.
By leveraging Claude:
- They created fake identities and application materials
- Wrote job cover letters and resumes
- Translated communications
- Developed and submitted code samples to pass hiring tests
Analysts say this tactic, long used by North Korean actors to bypass international sanctions and infiltrate foreign companies, is now being turbocharged by AI tools.
“Agentic AI can help them leap over cultural and technical barriers,” said Geoff White, co-host of The Lazarus Heist podcast. “Employers may unknowingly breach sanctions by hiring them.”
Industry Experts Call for Urgent Reform
While traditional cyber threats such as phishing emails and software vulnerabilities still dominate, experts warn that AI adds a dangerous new layer of sophistication.
“AI is a repository of sensitive information,” said Nivedita Murthy, a senior consultant at Black Duck Security. “It must be protected like any critical system.”
Anthropic has since reported the incidents to authorities and upgraded its detection and threat monitoring systems to prevent future abuse.
A New Era of AI-Driven Threats
These cases underscore the dark side of AI accessibility. As tools like Claude become more advanced and widespread, the line between innovation and exploitation continues to blur.
For businesses, governments, and cybersecurity professionals, the message is clear: AI is now part of the threat landscape, and proactive defence strategies are no longer optional—they’re essential.
For more tech news and insights, visit Rwanda Tech News, and explore similar topics and trends in the world of technology.
