Artificial intelligence is putting complex cyberattacks within reach of less experienced groups, Dell Technologies President and Chief Security Officer John Scimone said, raising the stakes for corporate defenders.
Anthropic reviewed 832 Claude accounts banned for cyber-related policy violations between March 2025 and March 2026. It found that 560 used its models for malware development. Of those same 832 accounts, 54 used the models for lateral movement, or moving through a compromised network.
Anthropic included accounts for which it had enough information to assess the techniques used. The counts describe activity in that selected group of banned accounts.
Scimone discussed the threat in an interview with theCUBE at the Dell Technologies New York Symposium on Tuesday. He said AI is reducing the money and specialist knowledge required to attempt sophisticated attacks, giving less capable actors access to more powerful methods.
An attacker’s level of expertise may therefore reveal less about what the attacker can attempt, Scimone said. He added that differences among hacktivists, criminal groups and nation-states still exist. He said security teams would need to use AI as they respond to threats involving the technology.
Dell described one defensive application in a June 5 statement. It said it was using Anthropic’s Claude Mythos Preview to identify and repair vulnerabilities in its products and newly written code. The company also said it was applying advanced AI models to protect its internal systems and networks.
Faster discovery and repair could help defenders and Dell customers, according to the company. Scimone’s warning points to the other side of that work: attackers with less expertise can also gain access to more capable tools.
The workload facing defenders was also raised in earlier EconoTimes coverage of AI-enabled cyberattack diversions. Cloudflare Asia-Pacific field chief information security officer Volker Rath told The Korea Herald that diversionary attacks were becoming harder to dismiss, adding to the alerts security teams must assess.
