- Anthropic claimed to have observed an AI cyberattack without substantial human intervention
- Experts say this claim is likely inflated by downplaying human intervention.
- The reports only describe what security professionals already know: AI tools speed up the attack process.
Anthropic recently reported that Chinese hackers had hijacked its Claude platform to launch fully AI-orchestrated cyberattacks, but this claim has since been met with skepticism in the cybersecurity community.
It seems likely that although AI carried out a significant portion of the attack (approximately 80-90%), the technology still needed vital human input: since AI cannot “think” for itself, it can only copy.
Some researchers believe this is just a marketing tactic to inflate the perceived capabilities of AI, or perhaps some scaremongering to fuel the narrative around the AI race between the United States and China.
nothing new
“I still refuse to believe that attackers are somehow capable of getting these models over obstacles that no one else can,” said Dan Tentler, executive founder of Phobos Group. Ars Technique.
“Why do models give these attackers what they want 90% of the time while the rest of us have to deal with ass-licking, evasion, and acid trips?”
While it may be true that AI has made leaps and bounds in recent months, it is still unlikely that it will be able to complete a number of complex tasks without human intervention. Tools are useful, but they enhance human capabilities rather than replacing them entirely.
“The implication here is that the attacker was using existing tools, but used an AI agent to take the place of the human who would normally handle those tools and go through the phases of the attack much faster,” said Tim Mitchell, senior security researcher in the Counterthreat Unit at Sophos X-Ops.
“From a defender’s perspective, that means there’s nothing new to defend against here, but the window to detect and defend against the attack is much narrower.”
Another point to note is that, according to Anthropic’s own reporting, only a “small number” of AI attempts to infiltrate organizations were successful, although it would have represented a first step in a rapidly evolving process.
TechRadar Pro He asked Anthropic for comment but heard nothing by the time of publication.
Follow TechRadar on Google News and add us as a preferred source to receive news, reviews and opinions from our experts in your feeds. Be sure to click the Follow button!
And of course you can also follow TechRadar on TikTok for news, reviews, unboxings in video form and receive regular updates from us on WhatsApp also.
The best identity theft protection for every budget
