Deepfake Awareness High at Orgs, But Cyber Defenses Badly Lag

AI-augmented deepfakes are increasingly present in cyberattacks on businesses and government bodies, with organisations broadly recognising the threat yet failing to match defence investments. On 7 October, OpenAI published research about adversaries using large language models to improve attack workflows, while a separate Ironscales survey on 9 October found that 85% of midsized firms had seen deepfake or AI-voice fraud attempts and 55% reported financial losses.

The study notes a preparation paradox: most organisations lag behind in investing in technical solutions even as they feel ready to respond, and two-thirds have not invested in AI-augmented threat defence. Among defenders, confidence remains high but actual protection is uneven, with many still reporting losses averaging around $167,000 in the past year.

Training is expanding, with 88% of organisations delivering deepfake-related training in the past year, up from 68% in 2024, while two-thirds view defence as a top priority for the next 18 months. Experts emphasise the need for policies and detection tools to reduce reliance on individual employees, given the growing sophistication of deepfakes.