“There’s no necessary governing mechanism over what it’s doing,” said Tom Barnett, senior director at legal innovation and technology advisory firm Maker5 and former chief data officer at Jackson Lewis. “It has the potential to go out into the world and start doing what you told it to do, but maybe you didn’t think of all the consequences of what you asked, or there’s some other bug in what it’s doing.”
Necessary Caution
Agentic tools can present organizations with a steep tradeoff between efficiency and safety. The ability to act autonomously is part of what allows AI agents to help users perform tasks more quickly, but that autonomy can be the source of damaging mistakes, especially if the system is not set up to allow users to trace and monitor the steps it takes.