Geoff Schomburgk discusses AI-driven cyber attacks and the gap in security practices in Yubico and Okta’s latest research
AI driven cyber attacks are becoming harder to detect, exposing gaps in business security practices, according to new research from Yubico and Okta.
Geoff Schomburgk, Regional Vice President at Yubico, highlights the growing risks and the need for stronger authentication as cyber threats evolve.
The 2026 Global State of Authentication report found that 42% of surveyed organisations experienced AI generated phishing attacks in the past year, while 37% of respondents struggled to distinguish AI generated emails from those written by humans.
Cyber criminals are using AI to create convincing phishing messages and deepfakes that impersonate managers, exploit trust and make fraudulent requests appear legitimate.
Despite growing awareness of these risks, businesses remain slow to adopt stronger security measures.
While 86% of respondents recognise passkeys as a safer alternative to passwords, fewer than half have adopted them.
The research also found that 72% of respondents prefer human oversight in decision making rather than relying entirely on autonomous AI systems.
The findings highlight the need for businesses to strengthen authentication, improve employee awareness and make secure practices easier to adopt as AI driven threats become increasingly sophisticated.
For more information, visit Yubico
