Google's Threat Intelligence Group (GTIG) said in a report published on November 16 that it expects fully autonomous attacks within three to six months, Infosecurity Magazine reported. Last quarter, GTIG saw a "paradigm shift" in how attackers use pen-testing frameworks, which are tools built to probe systems for weaknesses. The report names ARTEX and the open source tools CAIRN, CyberStrikeAI and PentAGI as increasingly popular with threat actors.
These tools run a set of cooperating sub-agents that plan, act and adapt across the stages of an intrusion. GTIG said they find easy-to-exploit flaws, use AI models to write single-use attack code on demand, and change tactics from live results. The methods stay the same: stolen credentials and unpatched systems. Speed and scale are what change. GTIG said attackers now trade stealth for speed, so attacks leave traces, but stealth will certainly increase. The article gives no figures on how many attacks use these tools.
Because the tools repeat known attacks so fast, slow patching and weak credential control cost more. Traces left today help defenders, but Google expects that to fade, so detection should not rely on noisy behaviour. Exabeam's Sally Vincent said watching for unfamiliar tool use, odd data access and shifts in activity volume can expose attacks.