F-Secure Corporation's Threat Intelligence Researcher Megan Squire @MeganSquire0 weighs in on the news that cyber criminals are using Claude AI to carry out sophisticated cyber attacks:
“Anthropic’s August Threat Intelligence Report confirmed that Claude not only performed ‘on-keyboard’ operations but also analyzed exfiltrated financial data to determine appropriate ransom amounts, and even generated convincing ransom notes that could be displayed on victim machines. The actor also leveraged Claude code to conduct automated reconnaissance. For example, the Claude code scanned thousands of VPN endpoints, identifying vulnerable systems with high success rates.
Why is this significant? Up to now, third parties have had to collect evidence of AI misuse from artifacts such as spam emails, news stories, and so on. There has been very little hard evidence of AI being used in the early stages of scam/fraud campaigns, for example to select victims, conduct reconnaissance, or communicate directly with victims. Mostly, AI has been used to enhance the quality of the bait. This report confirms that Claude, and likely other tools, are being used for nearly all stages of the scam kill chain.
While this confirms what we have long anticipated – the rise of AI tools used for cyber crime – it’s important we recognize that virtually any AI tool is susceptible to this kind of activity, and Anthropic’s transparency in this situation is the right approach.”
Read more here: www.bbc.com/news/articles/crr24eqnnq9o