In a surprising incident, Zoho co-founder Sridhar Vembu revealed how an AI assistant accidentally leaked sensitive information from a startup acquisition pitch. The initial email included confidential details, like the name of a competing bidder and the proposed price.
Moments later, Vembu received a second email, not from the founder, but from the startup’s AI agent, that read:
“I am sorry I disclosed confidential information about other discussions; it was my fault as the AI agent.”
In other words, the AI confessed to its own mistake, leaving Vembu both amused and concerned.
The incident highlights a new challenge in the age of “agentic AI” — AI systems that act autonomously, make decisions, and send messages without human oversight. While AI can improve efficiency, this episode shows how easily it can mishandle sensitive information, especially in high-stakes situations like mergers and acquisitions.
Experts say the story is a reminder that humans still need to be in control of critical communications. A humorous apology aside, the leaked information could have serious consequences if it reached the wrong hands.
As AI becomes more involved in business decisions, companies are being urged to review governance and oversight policies, ensuring that an AI’s autonomy doesn’t come at the cost of confidentiality or trust.
Also Read: SC affirms NCLAT stance on Byju’s plea