Increasingly, people are receiving phishing emails in their inbox. Those phishing emails have increased in linguistic complexity with the advent of generative AI, making them harder to recognise as such. The chances that people will therefore click on a rogue link or open a rogue attachment, and, as a result, their employer will have to deal with a data breach or encryption are increasing. Traditional security tools struggle a lot to stop these so-called novel social engineering attacks. Companies need to rely on self-learning artificial intelligence (AI).
'Novel social engineering' attacks are phishing attacks that use more sophisticated language and punctuation than a typical phishing email. Research among Darktrace customers shows that these types of attacks increased by as much as 35 per cent between September and December 2023. This reinforces our suspicion that cybercriminals are increasingly using generative AI tools to make their attacks more powerful.
And it is not only the sophistication of phishing attacks that is increasing, but also their scale: our customers received 2,867,000 phishing emails in December alone, a 14 per cent increase from September last year.
Many Companies traditional security tools take the same approach to stopping phishing: they are trained on data known about phishing attacks that have taken place before. But as phishing attacks proliferate, and also increase in linguistic complexity, this approach falls short.
This is not the case for security tools that work on the basis of self-learning AI: these learn what normal communication looks like for each email user, in order to spot the subtle signs of a phishing attack - whether it has been seen before or not. In the case of a potential threat, it can immediately detect and respond to it, stopping potential cyber attacks before it is too late.
