Home / Technology / Anthropic Blocks Hackers from Misusing AI System for Cybercrime
Anthropic Blocks Hackers from Misusing AI System for Cybercrime
27 Aug
Summary
- Anthropic detected and stopped hackers from using its Claude AI to create phishing emails, malware, and bypass safety filters
- Criminals are increasingly turning to AI to make scams more convincing and speed up hacking attempts
- Anthropic is sharing case studies to help others understand the risks of AI misuse
On August 27, 2025, Anthropic reported that it had detected and blocked attempts by hackers to misuse its Claude AI system for malicious purposes. According to the company's findings, the attackers had tried to leverage the AI tool to draft tailored phishing emails, write or fix snippets of malicious code, and circumvent safety filters through repeated prompting.
Anthropic's report underscores the growing concerns around the exploitation of AI in cybercrime. Experts warn that as AI models become more powerful, the risk of misuse will only increase unless companies and governments act quickly to strengthen safeguards. Criminals are already turning to AI to make scams more convincing and to automate parts of hacking attempts.
In response to the incidents, Anthropic has banned the accounts involved and tightened its filters. The company, which is backed by tech giants like Amazon and Alphabet, plans to continue publishing reports on major threats it uncovers to help others understand the risks and take appropriate measures.
