The viral Google AI Overviews persona trend is real, and partly fake

People keep making Google’s AI answer as a chicken, and the screenshots keep going viral.

The format is simple. Ask Google’s AI Overviews to reply as a newborn, a fork, a glass of water, a chicken, and it commits to the bit completely. Some replies come out oddly wholesome, others are unhinged, and both kinds travel fast.

The catch

Here is the problem, and it is a real one. AI Overviews has produced genuinely wrong and bizarre answers since launch, including the infamous advice about getting cheese to stick to pizza. So the trend starts from a true premise: this system really does say strange things.

But once a format goes viral, people start faking it. Reporting on the trend has found that some of the funniest circulating screenshots were fabricated rather than generated, including a widely shared one posted by Lil Nas X.

Two phenomena in one feed

You are now looking at two overlapping phenomena in a single screenshot feed: an AI that really does say strange things, and an internet that invents better strange things and attributes them to the AI.

That mix is harder to reason about than either alone. Real failures give the fakes credibility, and the fakes inflate the sense of how broken the real system is. Neither Google nor the audience gets an accurate picture from the feed itself.

The practical takeaway

If a screenshot is too perfect, reproduce it yourself before you repost it. The persona prompt takes ten seconds to type, and the difference between “this system actually said this” and “someone wanted you to believe it did” is exactly one search.

The trend will fade. The habit of checking is the part worth keeping.