Anthropic revealed it blocked hackers who tried to misuse its Claude AI system to create phishing emails, write malicious code, and bypass safety filters. The company’s report highlights growing concerns about AI being exploited for cybercrime and urges stronger safeguards.
The attackers attempted to use Claude for tailored scams, influence campaigns, and step-by-step hacking guides. Anthropic said it banned the accounts involved and tightened its filters, though it did not share technical details.
Experts warn that AI tools are making cyberattacks easier and more convincing. As models grow more advanced, the risk of misuse increases unless companies and governments act quickly.
Anthropic, backed by Amazon and Alphabet, said it maintains strict safety protocols and will continue publishing threat reports. Other tech giants like Microsoft, OpenAI, and Google have faced similar concerns, prompting regulatory efforts in the EU and voluntary safety commitments in the US.
Found this article interesting? Follow us on X(Twitter) ,Threads and FaceBook to read more exclusive content we post.
