Google’s AI tool ‘Backstory,’ which determines whether an image is AI-generated and investigates its origin, is speeding up media fact-checking.

Aug 24, 2026 08:00:00

In 2025, Google DeepMind announced ‘

Backstory ,’ an AI tool that, given an image and a text description, can determine if an image is AI-generated and investigate the context in which it has been used previously. Harvard University’s Nieman Journalism Lab reports that Backstory is accelerating media fact-checking.

Google’s new AI tool helps fact-checkers investigate AI fakes | Nieman Journalism Lab
https://www.niemanlab.org/2026/08/googles-new-ai-tool-helps-fact-checkers-investigate-ai-fakes/

Backstory is an AI tool developed by Google DeepMind that automatically checks whether an image was AI-generated after you upload it and enter basic information and answer some questions. As part of its AI detection capabilities, Backstory checks for SynthID, a watermark added by Google’s image generation AI Nano Banana, and C2PA metadata, which embeds content authentication information into images.

Furthermore, the AI agent compiles information such as whether there are any signs of tampering with the image, whether it has been previously published on the internet, and if so, in what context it was used, and then creates a report with citations. As of the time of writing, it is only available through the Trusted Testers Program, and thousands of journalists, OSINT (Open Source Intelligence) experts, and information literacy researchers are reportedly piloting it.

In fact, India Today, one of India’s largest news organizations, uses Backstory as part of its fact-checking team of six members to help identify a large amount of misinformation and false claims. Backstory is reportedly primarily responsible for the first stage of fact-checking and is said to be helping to speed up the process.

Bhar Krishna, who leads India Today’s fact-checking team, says, ‘Almost everyone on our team uses Backstory in combination with other tools. We use it as a quick reference because it provides background information and direction. The great thing about Backstory is that we can do all the work we used to do with five different tools and five different logins, all in one place.’

Previously, when fact-checking an image, it was necessary to first use an AI image detector to determine if it was AI-generated, and then perform an image search to find its origin. However, Backstory can perform all of these steps in one go.

Zoe Dalme, who worked on Backstory’s development at Google DeepMind, commented, ‘When people look at an image and try to understand its context, they tend to focus on only a part of the process. Backstory aims to integrate all of that into a single user experience.’

Dalme believes Backstory’s primary users are journalists, but sees that as it becomes more widely available in the future, there will be demand from a niche group that wants to know more detailed context, not just whether an image is AI-generated or not.

Digital literacy expert Mike Caulfield points out that verified information on the internet comes not only from experts like journalists and fact-checkers, but also from more casual amateurs. Support tools like Backstory can be a useful fact-checking tool for those who fall between experts and amateurs.

Caulfield argues that Backstory’s most useful feature isn’t its ability to identify whether an image is AI-generated or not, but its ability to track its online history. ‘Getting to know the whole picture—the whole story behind it—was actually the most time-consuming part. With Backstory, we can reduce that initial process, which usually takes 50 minutes, to just 3 minutes,’ Caulfield said.

Like many other AI models, Backstory is not perfect, and a disclaimer appears at the bottom of the tool stating, ‘Backstory can make mistakes, so please be sure to check it carefully.’ In fact, a reporter from Nieman Journalism Lab reported that when they tried Backstory with their own photo, it mistakenly identified them as another journalist.

Furthermore, Backstory can only reference images that are indexed by search engines when tracking the context of images. Therefore, it may not be able to track images originating from places that are not indexed, such as social media platforms like TikTok, or private groups on WhatsApp or Signal.

According to Krishna, India Today does not use reports generated by Backstory directly in its fact-checking articles; the information obtained is always manually verified by the journalists. Nevertheless, Krishna claims that Backstory is the first image verification tool for fact-checkers developed by a major technology company and helps streamline daily operations.