Google to Flag Fake News Photos

Google is adding new tools to make it easier to check if an image is credible and as it appears. But the move has highlighted the irony of Google's own involvement in "generative" artificial intelligence.

The tools include several additions to Google's own search results, along with a dedicated tool for checking any image found online.

The search tools have been in testing and are now rolling out to all users, starting with English language searches in the US. They are wrapped up in a featured called "About this image".

This will accompany images with information including if and when they have previously been spotted by Google search on other websites. Google says this will make it easier to spot if an image is being used out of context, for example if a photo from a past news event is being falsely presented as part of a current situation. (Source: blog.google)

Separate Fact Check Tool

There will also be a summary of how the image is being used across different web pages, particularly news and fact-checking sites. This could highlight where people have drawn attention to whether an image is credible or not.

Finally, Google will show metadata that could show when and where an image was taken. However, this will rely on publishers having added such data (or failed to remove it), so will be quite hit and miss.

As most users will encounter images in contexts other than a Google search, the company is also adding a dedicated "Fact Check Explorer" where users can either upload an image or input its URL. Google will then show whether or not the image has ever been featured by a fact checking service.

Even when this isn't the case, users will see more context about where the image has appeared before, including cases where it may have been cropped differently.

AI Problematic

One problem is that such tools are generally based around the idea that an image is indeed a photograph, even when it's presented in a misleading context or false described. WIRED notes this won't always help with images which look like photographs but have been created by artificial intelligence tools. (Source: wired.com)

Indeed, Google even offers such tools itself. For its part, it says any images the tools create will carry a metadata label that identifies it as an AI creation. It hopes - perhaps naively - that other AI developers will follow suit.

What's Your Opinion?

Do you ever check if an image is "genuine" and if so, how? Is trying to show the correct origin and context of images a worthy effort or a lost cause? Do you think more people will be fooled by AI images presented as news photographs in the future?