10/15/2026
๐ฐ๐ต.๐ด% ๐ฎ๐ฟ๐ฒ ๐๐ฒ๐ฒ๐ถ๐ป๐ด ๐ฟ๐ฒ๐ฑ. ๐๐ ๐๐ผ๐๐ฟ ๐ผ๐ฟ๐ด๐ฎ๐ป๐ถ๐๐ฎ๐๐ถ๐ผ๐ป ๐ผ๐ป๐ฒ ๐ผ๐ณ ๐๐ต๐ฒ๐บ?
Salt Securityโs latest research found that 49.8% of respondents reported or suspected an AI agent had taken an unintended or unauthorized action in the past year.
The unsettling part? If you don't know what your AI agents are doing, you might not even know there's a problem.
As AI agents move from answering questions to taking action across enterprise systems, ๐ป๐ผ๐ ๐ธ๐ป๐ผ๐๐ถ๐ป๐ด ๐ถ๐๐ป'๐ ๐๐ต๐ฒ ๐๐ฎ๐บ๐ฒ ๐ฎ๐ ๐ฏ๐ฒ๐ถ๐ป๐ด ๐๐ฎ๐ณ๐ฒ.
๐๐ ๐๐ผ๐๐ฟ ๐ผ๐ฟ๐ด๐ฎ๐ป๐ถ๐๐ฎ๐๐ถ๐ผ๐ป ๐ผ๐ป ๐๐ต๐ฒ ๐ฟ๐ฒ๐ฑ ๐๐ถ๐ฑ๐ฒ, ๐ผ๐ฟ ๐ฑ๐ผ๐ฒ๐ ๐ถ๐ ๐ป๐ผ๐ ๐ธ๐ป๐ผ๐ ๐๐ฒ๐?