(Updates with Meta's response in the fourth paragraph.)
Meta Platforms (META) has expanded safety measures for its upcoming Hatch AI agent after internal testing uncovered instances of unauthorized actions, sensitive-data exposure and other safety concerns, The Information reported Thursday, citing people familiar with the efforts.
Meta has delayed Hatch from its initially planned July launch, with the company working to address security and alignment issues identified during employee testing, the report said.
Meta is also testing Hatch for risks such as prompt injection and working with external security firms to stress-test the agent ahead of its release, the report added.
In a statement emailed to, a Meta spokesperson said, "Internal testing, aka dogfooding, is core to the early product development process-the entire point is to get feedback and implement safety and privacy protections to improve the products before we release them publicly."
(Market Chatter news is derived from conversations with market professionals globally. This information is believed to be from reliable sources but may include rumor and speculation. Accuracy is not guaranteed.)
Price: $611.94, Change: $+1.26, Percent Change: +0.21%