AI set to play key role in future phishing attacks

A staggering increase in QR code phishing (quishing) attacks during 2023 saw them skyrocket up the list of concerns for cyber teams globally, according to Egress.

Attacks were both prolific and highly successful, demonstrating how cybercriminals effectively combine available technology with consumer familiarity (or complacency) at scale.

The evolution of phishing attack payloads

In 2021 and 2022, QR code payloads in phishing emails were relatively rare – accounting for 0.8% and 1.4% of attacks respectively. In 2023, this jumped to 12.4% and has continued at 10.8% for 2024 so far.

Social engineering has also increased, now representing 19% of phishing attacks and phishing emails are over three times longer than they were in 2021, likely due to the increase in use of generative AI. On the other hand, the use of attachment-based payloads has decreased since 2021; three years ago, these accounted for 72.7% of attacks detected by Egress, and by the first quarter of 2024, this had fallen to 35.7% as threat actors evolve their payloads to evade cybersecurity efforts.

Following initial phishing email attacks, Microsoft Teams and Slack account for 50% of second steps in multi-channel attacks, and the Egress team only expects this to rise in popularity amongst cybercriminals. Microsoft Teams was the most popular second step in multi-channel attacks, accounting for 30.8%, followed by Slack (19.2%), and SMS (18.6%).

With security awareness training (SAT) generally focusing heavily on educating employees about email-based attacks, and a perceived legitimacy with these messaging channels, it’s no surprise that Microsoft Teams experienced a 104.4% increase in 2024 compared to the last three months of 2023.

Cybercriminals use AI for deepfakes, automated phishing

Deepfakes continue to hit the headlines, and the use of Zoom and mobile phone calls as the second step in multi-channel attacks has increased in the first quarter of 2024 compared with the last quarter of 2023; Zoom by 33.3% and mobile phone calls by 31.3%. Researchers predict the use of video and audio deepfakes in cyberattacks will increase over the next 12 months and beyond.

Generative AI is also expected to increase attack success rate, including creating payloads such as malware, phishing websites and invoices for wire fraud attacks as cybercriminals look to streamline their processes and deliver more efficient campaigns at even swifter pace.

The report reveals that in the first three months of 2024, there was a 52.2% increase in the number of attacks that got through SEG detection. 68.4% of these attacks passed authentication checks, including DMARC, which is a primary detection capability used by SEGs.

Unlike integrated cloud email security (ICES) solutions, SEGs are less effective against legitimate but compromised third-party accounts, which is where most of these attacks have been sent from. Sitting at the network’s edge, SEGs utilize definitions libraries and scan for known threats using signature-based and reputational-based detection, with this detection mechanism remaining relatively static despite the rapid evolution of phishing threats.

Obfuscation techniques frequently bypass SEGs such as hijacking legitimate hyperlinks and masking hyperlinks to phishing websites within image-based attachments like JPEGs. These two techniques make up 45.5% of obfuscation methods that bypass SEGs, and layering multiple techniques is increasingly popular for avoiding detection.

Millennials are the top targets for phishing attacks

The report reveals that Millennials are the top targets for phishing attacks, receiving 37.5% of phishing emails. The most targeted industries are finance, legal and healthcare, with people working in accounting and finance teams receiving the most phishing emails, followed by marketing and HR. Unsurprisingly, the most targeted job role is the CEO and 13.4% of phishing attacks impersonated someone the victim knew such as CEOs and senior leadership.

Social engineering is evident in the most phished day of the year so far, as February 9th came out on top in the lead up to Valentine’s Day. Utilizing a widely celebrated holiday to personalize phishing attacks has always been popular, but the rise of AI will lead to these being increasingly convincing.

“The third edition of the Egress Phishing Threat Trends Report is jam packed with crucial themes and predictions for the threat landscape for 2024. We look at hot topics that have dominated headlines, including the rise of QR phishing and AI-powered attacks, plus we analyze the ways cybercriminals are engineering attacks to get through detection by secure email gateways,” said Jack Chapman, SVP of Threat Intelligence at Egress.

“The one thing that won’t change in 2024 is cybercriminals investing heavily in attacks that give them the highest rewards. Some tactics will stay the same, but where returns diminish or disappear entirely, new tactics will emerge. Looking at the trends explored in the latest report, we can say with certainty that AI-powered attacks are here to stay, and our Threat Intelligence team predicts AI will be used in some way in every phishing attack in the next 12 months, leading to lucrative paydays for cybercriminals,” Chapman concluded.