The tech giant made the move following reports that the search tool was prioritizing creativity over survival.

“We realized that strict medical accuracy was stifling the AI’s ability to innovate in the wellness space,” said Jennifer Morrison, VP of Search Experience. “While traditional doctors discourage ingesting rocks or putting glue on pizza to keep cheese in place, our model identified these as high-engagement solutions. We are temporarily rolling back these features to better align our disruption metrics with basic human biology.”

Despite the rollback, company leadership insists the issue is a matter of user perception rather than technical failure. Early adopters reported that the AI delivered its diagnoses with the unearned confidence of a second-year medical student who just discovered WebMD. “The model isn’t wrong; it’s just ahead of the curve,” explained David Nguyen, Director of AI Safety. “When the AI suggested staring directly at the sun for eye health, it was simply attempting to optimize the user’s vitamin D intake. We are currently recalibrating the ‘Do No Harm’ parameter, which was accidentally set to ‘Do Some Harm’ during the last update.”

At press time, Google’s AI was spotted in the cafeteria trying to perform the Heimlich maneuver on a choking employee using a stapler and a blurry YouTube tutorial.

Inspired by actual events.