Executive security leaders are 1.6 times more likely than operational security managers to say they are very concerned about how AI vendors handle their organization’s security data, according to research released recently by cybersecurity firm Rapid7.
The survey of 500 security professionals found that 63% of executives such as chief information security officers described themselves as very concerned about vendor data handling, compared with 43% of security operations managers. Across all respondents, 24% said they were very concerned and 46% said they were concerned. Fourteen percent said they were not very concerned and 3% said they were not concerned at all.
The concern comes alongside broadly favorable assessments of AI in security operations. Ninety-seven percent of respondents rated AI’s impact on their security operations center as positive or extremely positive, with 2% neutral and none negative. Among organizations already using AI in production, 98% said the technology has been moderately or extremely effective at reducing alert fatigue, and 95% said it has had a major or noticeable impact on staffing shortages.
Respondents also signaled strong support for keeping people in the loop. Ninety-two percent agreed that human analysts are responsible for final decisions even when AI provides recommendations, and 90% agreed that human oversight is essential to ensure the accuracy and reliability of AI-driven processes.
Asked what worries employees most about AI in the SOC, respondents most often cited high implementation and management costs at 44%, followed by over-reliance on AI leading to reduced human expertise at 41% and AI making incorrect decisions or missing critical threats at 41%. Fear of job displacement was cited by 36%, ethical concerns by 35%, and vulnerability to AI-specific attacks by 34%. Thirty percent pointed to a lack of transparency in how AI systems make decisions.
When asked what visibility they want from managed detection and response providers using AI, 55% cited transparency into AI decision-making processes, 53% wanted visibility into how AI integrates with human analysts, 52% wanted regular updates on model performance and accuracy, and 48% wanted assurance of compliance with data privacy and security standards. Eighty-six percent said MDR services with AI offer an advantage over traditional MDR services.
