Security teams rank employee AI data use above AI-powered attacks
Original title: Security pros reckon the biggest AI risk comes with an employee badge
In O’Reilly’s September survey, 55% had seen AI-generated phishing, but only about a quarter called it a top concern.
Help Net Security reports an O’Reilly survey of security professionals who worry more about staff pasting company data into AI tools than about AI-powered attacks. Data leaks through commercial AI tools and trust in unchecked AI output each ranked among the main concerns for 49%. AI-generated phishing was the attack most often encountered, while prompt injection and loosely permissioned AI agents drew less concern. Mike Loukides said vendor promises not to train on enterprise prompts deserve skepticism and that weak AI-use rules track with more shadow AI.
Unconfirmed: Findings reflect O’Reilly’s survey and commentary by Mike Loukides, not confirmed incidents.
- O’Reilly surveyed 209 security professionals in September.
- Over half said coworkers at least occasionally put sensitive data into unapproved AI tools.
- Widespread shadow AI was 17% under comprehensive policies and 35% under basic ones.
- Prompt injection worried 11% of participants.
- Two-thirds use chatbots for research; about a quarter use AI for detection or response.
- Organizations
- O'Reilly
From the original article
3 min read at helpnetsecurity.com
Employees pasting company data into AI tools worry security teams more than AI-powered attacks do. Most-concerning risks of AI agents (Source: O’Reilly) AI-generated phishing, the attack they run into most often, made the list of top concerns for only a quarter of the 209 security professionals O’Reilly surveyed in September.
By Sinisa Markovic · helpnetsecurity.com
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.