Advances in AI are resulting in the development of autonomous agents that have the ability to plan, reason and take actions to fulfil a specific goal, which is making them increasingly attractive to hackers.
Agents could be used to identify and target vulnerable computer systems and steal information at scale, but they could also be used to protect systems too.
While the use of AI agents to conduct cyberattacks is still in its infancy, experts believe they will become more widespread in the next few years.
However, one of the problems with autonomous agent hacking is how to detect them in the first place, which is why AI research organisation, Palisade Research, has created a system called LLM Agent Honeypot that tries to catch AI agents attempting to hack into systems.
The system has so far logged 11million attempts to access it, most of which have been from humans and simple bots, but among these, the researchers have detected eight potential AI agents, two of which have been confirmed as such and appear to have originated in Hong Kong and Singapore.