Researchers are warning that AI threatens to increase the sophistication and effectiveness of cyberattacks, according to a recent blog post by the Wall Street Journal.

Earlier this year, cybersecurity firm Darktrace found a brand new type of cyberattack at a company in India. The software used rudimentary machine learning to observe and learn patterns of normal user behaviour inside a network. The malware could learn as it spread, altering its methods to avoid detection for as long as possible.

Darktrace’s researchers stopped short of describing the malware as fully-fledged AI-driven software, however did say that the malware’s methods were “early indicators” of AI.

It is easy to see how the learning algorithms that we associate with AI can be used for optimistic attacks. Worse, big data and large numbers of permutations are the sweet spots for AI. We know there is already plenty of processor power devoted to cyberattacks. The next step to the war of the machines is inevitable.

Harry Crawford contributed to this post.