Google's new tools to spot AI fakes

With tools like Midjourney pumping out photorealistic images at the drop of a prompt, seeing is officially no longer believing. As generative tech spreads, distinguishing a camera snap from synthetic fiction has become a genuine headache.

Google is rolling out two new features aimed at giving users the upper hand against deceptive imagery.

Tracing an image back to its roots

The first weapon in Google's arsenal is a tool called "About this image," set to roll out in the US over the coming months.

Users can access it by clicking the three dots next to a result in Google Image Search, scanning with Google Lens, or swiping up in the Google App. Once opened, it delivers critical context in three clear buckets.

First, it reveals when Google first indexed the picture or similar visuals, telling you whether an alleged breaking photo has actually been floating around for years.

Second, it flags where the image first surfaced on the web, helping you tell if it originated on a credible news wire or an obscure forum.

Third, it lists where else the image has been published, including coverage by fact-checking organizations, newsrooms, and social media platforms, making it easier to spot twisted narratives.

Watermarking the source files

Detecting fakes is only half the battle; tagging them at birth is the other.

Google pledged to embed metadata markups directly into its own generated image files, officially labeling them as AI-produced.

Crucially, Google confirmed it has already secured buy-in from Midjourney and Shutterstock to adopt these markups, with more industry players expected to follow suit.

Google Search product manager Cory Dunton pointed to a 2022 Poynter study showing that 62 percent of people believe they encounter misinformation at least weekly, underscoring the urgency of integrating verification tools directly into everyday search workflows.

Why it matters

For everyday internet users and researchers, having instant access to an image's origin and index history inside Google Lens removes the heavy friction of manual reverse-image investigations. It is practical and worth tapping into the moment a viral photo looks suspicious.

Industry-wide, Google getting Midjourney and Shutterstock onboard with standardized metadata is a smart, collaborative move. Tackling synthetic media cannot happen in isolation inside search engines; it requires the generators and the stock libraries to play along.

However, we think the fundamental flaw here lies in user intent. Both tools require someone to already harbor enough skepticism to pause, click the menu, and investigate. The users most vulnerable to outrage-inducing fake photos are precisely the ones least likely to inspect the metadata. While providing context is undeniably a step in the right direction, it feels like a soft brake against an incoming freight train of synthetic media.