Nearly All Security Leaders Are Bracing for AI-Driven Attacks, MIT Technology Review Study Finds
A Darktrace-backed MIT Technology Review Insights survey of 300+ executives finds 96% preparing for AI-powered attacks and many turning to AI defences.
Security leaders expect attackers to start using artificial intelligence, and almost all of them are already getting ready for it. That is the headline of a study by MIT Technology Review Insights, publicised by the autonomous cybersecurity AI company Darktrace, which found that 96% of respondents are preparing for AI-powered cyber attacks. Many are adopting AI-based defences of their own.
The researchers questioned more than 300 managers, directors and C-level executives around the world between December 2020 and January 2021, asking how they deal with today's cyber threats and those they expect next. The main results:
- Email attacks worried respondents most (74%), followed by ransomware (73%) and cloud-based attacks (68%).
- 96% are getting ready for the arrival of AI attacks.
- 68% think AI will be used for impersonation and spear-phishing.
- 60% believe responses led by people cannot keep pace with automated attacks.
The report features a case study on McLaren Racing. The Formula 1 team relies on Darktrace's Autonomous Cyber AI to spot and react to cyber attacks that move quickly, and the study describes one occasion when the AI halted a sophisticated email impersonation attack in the middle of a busy race weekend.
It also draws on Max Heinemeyer, Darktrace's director of threat hunting and a former white hat hacker. He looks at the problems posed by "fearware": scam emails, tightly targeted, that played on public anxiety about the pandemic.
Laurel Ruma, US editorial director at MIT Technology Review, noted that 60% of the people surveyed were C-level executives and directors. In her reading, the results make plain that cybersecurity is a serious, concrete concern for business leaders, and that AI will have a very large role in protecting every enterprise.
Nicole Eagan, Darktrace's chief strategy and AI officer, described the findings as evidence of "a watershed moment", with executives getting ready for a new era of attacks. She argued that defences built on studying past attacks will be poorly equipped against offensive AI, and that human security teams will need support from a fundamentally new approach based on self-learning technology and autonomous response.