AI could soon defeat most cybersecurity measures

Three in four (76%) cybersecurity professionals fear that there will soon be a malicious AI capable of bypassing most known cybersecurity measures, according to a new report from Enea.

The report found that 26% expect such an AI to be uncovered in the next year and 50% consider it likely in the next five years.

Cybersecurity professionals believe that common cyberthreats, such as phishing, social engineering and malware are among those most likely to become more dangerous due to AI.

In addition to concerns over offensive AI outpacing defensive AI, 77% of cybersecurity professionals are seriously concerned that a rogue AI will emerge with behaviour that veers away from its intended purpose, leading to unpredictable and dangerous behaviour.

But AI is also anticipated to bolster threat detection and vulnerability assessments, with deep learning for detecting malware in encrypted traffic seen as holding the most promise. As a result, 48% of surveyed professionals anticipate a positive impact from AI, and 72% believe AI automation will play a key role in alleviating cybersecurity talent shortages.

A majority (61%) of organisations have yet to deploy AI in any meaningful way as part of their cybersecurity strategy, but 41% consider AI to be a top priority.

Enea industry analyst Laura Wilber Sr said understanding the profound impact of AI on cybersecurity is crucial for navigating the evolving threat landscape.

“That begins by listening closely to the concerns and hopes of cybersecurity leaders and their teams on the frontlines,” she said. “This report confirms growing concerns around the malicious use of AI, but it also highlights some remarkable innovations in the use of AI to streamline and automate defences.”

Image credit: iStock.com/Vertigo3d