
AI fuels rise in complex social engineering attacks worldwide
LevelBlue has released new research indicating that artificial intelligence is increasing the complexity of social engineering attacks and exposing organisational vulnerabilities.
The report, titled Data Accelerator: Social Engineering and the Human Element, highlights that human behaviour remains a significant risk factor for many companies. As AI accelerates the development of more convincing and scalable attack methods, organisations face increasing challenges in protecting themselves against cyber threats.
According to the study, 41% of organisations reported a significantly higher volume of cyberattacks over the past year. The research captures a growing awareness of the dangers posed by social engineering, including new methods that leverage AI to manipulate or deceive individuals within businesses.
Despite this awareness, many organisations are not adequately prepared. The findings show that 44% of respondents believe an AI-powered attack is likely within the next year, but only 29% feel prepared to face such an incident. Furthermore, just 20% of all organisations described their defences against AI-enabled cyber adversaries as highly effective.
Specifically, 38% acknowledged underpreparedness for AI-driven social engineering threats: these include automated attacks, deepfake-based videos, and voice scams. When asked about their readiness to counter specific threats, survey respondents identified differing levels of preparedness: 57% for personal information exfiltration, 56% for business email compromise, followed by lower levels for other social engineering threats such as phishing (51%), smishing (43%), quishing (41%), insider threats or account takeover (44%), and deepfake or synthetic identity attacks (32%).
Employee awareness and training present significant obstacles to mounting effective defences against social engineering attacks. The research found that 59% of organisations believe their employees are finding it increasingly difficult to distinguish between genuine and fraudulent communications. Despite this trend, only 20% of organisations expressed confidence in their current efforts to educate employees, and a mere 32% have engaged training and awareness experts over the past year.
Investment priorities also illustrate a shift in strategy. Rather than focusing primarily on staff training, 33% of organisations reported significant investment in cyber resilience processes, and 31% allocated resources specifically to generative AI tools intended to help defend against social engineering attacks.
Geographically, Europe was identified as the most prepared region for AI-driven attacks, with 66% of organisations there expressing readiness. The report also notes that Europe is ahead in prioritising employee education to enhance cyber resilience compared to other regions.
"Establishing a culture of cyber resilience is imperative for organisations to effectively prepare for the emergence of more sophisticated social engineering attacks," said Theresa Lanowitz, Chief Evangelist of LevelBlue. "These attacks exploit human behaviour, so without the proper investment into education and training, including cyber resilience processes and engaging cybersecurity consultants, organisations and their employees remain vulnerable."
While investment in generative AI as a defensive measure is becoming more common, only 24% of organisations rated themselves as highly effective in deploying and using AI to enhance security. Additional defensive measures such as Zero Trust Architecture are less widely adopted, with only 13% of respondents indicating major investment in this area to reduce the risks related to compromised credentials and social engineering.
Based on its findings, LevelBlue recommends four key actions for organisations aiming to improve their resistance to social engineering threats: increasing leadership engagement in building a cyber resilient culture, implementing regular cybersecurity training to keep staff informed about emerging threats, evaluating investments with attention to both human and technical vulnerabilities, and engaging external providers for enhanced defence capabilities.
The research methodology involved a quantitative survey, conducted by FT Longitude in January 2025, that included 1,500 C-suite and senior executives across 14 countries and seven sectors including energy and utilities, financial services, healthcare, manufacturing, retail, transportation, and US state and local government, as well as higher education.