Artificial intelligence is making cyberattacks even more dangerous, according to a new report published by Anthropic. The AI company analysed 832 AI-enabled cyberthreat cases from March 2025 to March 2026. In 67 percent of those, the hackers used AI for attack preparation.
Attack preparation can come in many forms, such as writing malware and disrupting system performance. “Malicious actors are using AI in ways that make them more dangerous. More specifically, threat actors are using AI in the later, more complex stages of their cyber operations,” Anthropic said in its report.
According to the company, 6.5 percent of threat actors used AI to assist with “lateral movement,” which involves navigating deep inside a compromised network. In Anthropic’s study, researchers found signs that AI may be helping make attacks more dangerous. In the first six months of the research, 33 percent of attackers were rated at a medium risk or higher. Over the next six months, that number rose to 56 percent, approximately a 70 percent increase.
“Now that AI can perform highly technical tasks on an actor’s behalf, there’s little correlation between the skill of a threat actor and how many techniques they use,” the company stated.
Threats to come
In order to combat the issue, Anthropic is committing to adding safeguards. “The findings from this analysis helped inform the safeguards we build into our models. For example, we’ve developed and deployed cyber safeguards on our most capable models to detect and block some of the activities uncovered here, like developing malware or mass data exfiltration,” the report says.
In September 2025, Anthropic detected suspicious activity stemming from a highly sophisticated espionage campaign by a Chinese state-sponsored group. Using Claude Code, the hackers infiltrated roughly 30 global targets. Ultimately, it was disrupted by Anthropic before the situation worsened.
Mapping it against the Mitre Att&ck framework, Anthropic determined that the hackers in the case were medium-risk actors.
Verizon’s 2026 Data Breach Report notes that 31 percent of breaches start with software vulnerabilities, nearly 50 percent of all breaches now involve ransomware, and mobile devices are the new favourite target for AI-driven cyberattacks.
“Threat actors are using AI to work faster at every stage–from spotting security gaps to writing malware,” Anthropic’s report says. As AI-driven cyber threats continue, Anthropic is reportedly in talks with Mire to continue observing attack frameworks while also updating security and creating new defensive strategies. – Inc/TNS
