A recent report by AI security vendor Reco reveals that 80% of AI tools used in large enterprises operate without IT oversight, increasing operational risks for organizations. The study highlighted that SMBs have approximately 414 unsanctioned AI tools per 1000 employees. The report further indicated that many AI systems can access sensitive files and execute commands, heightening the threat to data security.
Over the past 18 months, 637 vulnerabilities have been detected in AI agents, with an alarming increase in disclosures, outpacing the ability to patch them. The study emphasized the need for enhanced governance and security measures in AI integrations.