Ransomware Expert Warns AI Will Transform Attacks in 2027
Ransomware is set to be revolutionised in 2027 by artificial intelligence, according to Bob McArdle, TrendAI's director of cybercrime research. Speaking at the IRISSCON 2026 conference in Dublin on 11 November, he described the past couple of years as "pretty stagnant and boring" for researchers. So far, criminals have mainly used AI to work faster, such as finding the most valuable data once inside a network, and in some cases using chatbots to handle early ransom negotiations. The basic attack process, including the use of initial access brokers and double extortion, has not changed.
McArdle warned that this is "a drop in the ocean" compared with what is coming. He pointed to AI frameworks like Artex, which lower the barrier to finding and exploiting weaknesses quickly. In his words, anything connected to the internet that is misconfigured or vulnerable will be compromised. He expects this to devalue the market for initial access, pushing brokers to offer privilege escalation to ransomware groups instead. Once inside, AI tools and agents could shrink the attack timeline to machine speed.
He also highlighted a risk from organisations centralising their data to support AI processing. While efficient for business, it gives attackers new opportunities, including threatening to tamper with the AI orchestration around an organisation's data as a further extortion tactic.