Experts now warn that the upcoming surge of AI-driven cyberattacks will outpace those carried out by human hackers in the past.
This is the warning conveyed by AI firm Anthropic in a recently leaked blog post, where it warned that its upcoming AI model, named Mythos, along with similar models, has the capability to exploit vulnerabilities at an unparalleled speed. Moreover, it is not alone in this concern.
OpenAI indicated in December that its forthcoming models present a “high” cybersecurity threat.
Experts have already noted that AI can worsen existing risks and swiftly create new software vulnerabilities.
However, the emergence of AI agents, or AI assistants capable of performing tasks independently, increases this risk to a new dimension, according to some experts.
Next wave of AI cyberattacks may outpace human hackers
A single AI agent could identify vulnerabilities and potentially exploit them more quickly and persistently than numerous human hackers.
“The agentic attackers are on the horizon,” stated Shlomo Kramer, founder and CEO of the cybersecurity and networking firm Cato Networks.
“This marks a pivotal moment in the evolution of cybersecurity”, he added.
“Although Mythos is currently far ahead of any other AI model in cyber capabilities, it presages an upcoming wave of models that can exploit vulnerabilities in ways that far outpace the efforts of defenders,” Anthropic stated in a draft.
The company is allowing select organisations to test the model in advance to enhance their systems “against the looming wave of AI-driven exploits,” it noted.
Additionally, Anthropic is secretly alerting government officials about the potential for extensive cyberattacks facilitated by Mythos, as reported by Axios.
Read more: New study reveals AI models can now control their own behaviour
