Matthew Rosenquist predicts AI will supercharge social engineering attacks in 2026, making them far more convincing through personalized messages, deepfake voices, and fake websites. Success rates could jump from 30% to 50%, with most malicious emails slipping past current filters unless defenses improve fast.
Chapters
Social engineering will skyrocket in 2026 as attackers fully embrace AI to significantly increase the number of attacks and their believability. Cybersecurity will need to adapt to the dramatic improvements by attackers who use highly automated social engineering tools to victimize their targets. Attacks via email, phone calls, text messaging, and on social platforms will be amplified through the use of AI. AI will orchestrate attacks that will be personalized, professionally written, culturally fluent, and emotionally compelling. Attackers will incorporate victim information from public sites and dark web data to make messages more personalized, urgent, and realistic, resulting in greater effectiveness at compelling them to engage click a malicious link, install malware, share sensitive data, grant access, or other nefarious acts. Now imagine seeing an email that impersonates a coworker, an old friend, or someone you just recently met at a conference, which praises your comments on a topic, discusses issues specific to your work division, recalls a past event, or weaves a narrative about your favorite hobby, all in wording, dialect, and an accent which seems natural, familiar, and trustworthy. By the end of the year, cybercriminals will impersonate other people and entities with amazing accuracy and believability. Their AI tools will create fictitious personalities, websites, images, voice, and highly engaging social media accounts, all to fool and manipulate victims. Expect an increased rate of social engineering success, jumping from 30 to 50% over what attackers were able to accomplish in 2025, and a spike in the number of messages which bypass current filters. This will become the norm until cybersecurity is able to adapt with better tools, processes, and training. It is safe to say that a lot of people will fall victim to AI-enhanced automated social engineering attacks in 2026.
