Meta is modifying its AI's suggested prompts after a viral video showed the feature posing invasive personal questions about a user's young daughters. The user, Kalie Robins, posted a video of herself and her child, which was cross-posted to Facebook. Afterward, Meta AI presented a prompt asking to identify the child.
Upon engaging with the prompt, the AI reportedly aggregated detailed information about both of her daughters by referencing years of posts from her and her relatives, including some she had deleted. A Meta spokesperson acknowledged the company "missed the mark" and stated the feature should never have generated such questions. The company says it has now implemented a fix to prevent the AI from suggesting prompts on personal topics.