AI Cyber Risk: Social Engineering Losses Soar 85%

Artificial intelligence is reshaping cyber risk, but not in the way many might expect. Claims data shows that AI's biggest impact isn't from new, AI-specific attacks. Instead, AI is making familiar tactics like social engineering more effective and financially damaging. Resilience's 2026 Midyear Cyber Risk Report reveals no incurred losses from AI-native attacks in the first half of 2026. What's interesting is how AI acts as a force multiplier for human-enabled attacks. Social engineering accounted for 85% of incurred losses, a significant jump from 17% just a couple of years prior. Payment fraud losses also tripled. The bottom line is that the targets remain people, credentials, and internet-facing devices. This information helps organizations prioritize spending based on actual financial impact, rather than perceived risks.