Criminals increasingly turn to artificial intelligence to make scams more convincing and cyberattacks faster.
Anthropic, an AI company, announced that it had successfully detected and blocked attempts by hackers to exploit its AI system, Claude. The attackers aimed to use Claude to write phishing emails, generate malicious code, and bypass safety filters.
The company’s report highlighted growing concerns over the use of AI tools in cybercrime, prompting calls for stronger protective measures from both tech companies and regulatory bodies. According to Anthropic, the company’s internal systems prevented these attacks, and case studies will be published to shed light on the incidents.
The report detailed attempts that included crafting targeted phishing emails, writing or modifying sections of malware code, and circumventing restrictions through repeated requests. It also mentioned efforts to launch intrusion campaigns by generating persuasive messages at scale and providing step-by-step guidance to inexperienced hackers.
Supported by companies such as Amazon and Alphabet, Anthropic did not disclose technical details like IP addresses or the content of the requests but confirmed that associated accounts were blocked and that its filters were tightened following the detection of these activities.
Experts warn that criminals are increasingly leveraging AI to make scams appear more realistic and to accelerate cyberattacks. AI tools can write convincing phishing messages, automate portions of malware development, and even assist in planning attacks.
Security researchers stress that as AI models become more powerful, the risk of misuse grows—unless companies and governments act quickly. Anthropic says it has strengthened its safety guidelines, while other tech giants such as Microsoft, OpenAI, and Google face similar concerns regarding potential misuse of their AI systems.
Meanwhile, governments are moving toward regulation. The European Union is focusing on the “AI Act,” while the United States emphasizes voluntary safety commitments from major developers.
Support Dawat Media Center
If there were ever a time to join us, it is now. Every contribution, however big or small, powers our journalism and sustains our future. Support the Dawat Media Center from as little as $/€10 – it only takes a minute. If you can, please consider supporting us with a regular amount each month. Thank you
DNB Bank AC # 0530 2294668
Account for international payments: NO15 0530 2294 668
Vipps: #557320
Support Dawat Media Center
If there were ever a time to join us, it is now. Every contribution, however big or small, powers our journalism and sustains our future. Support the Dawat Media Center from as little as $/€10 – it only takes a minute. If you can, please consider supporting us with a regular amount each month. Thank you
DNB Bank AC # 0530 2294668
Account for international payments: NO15 0530 2294 668
Vipps: #557320
Comments are closed.