In the days following Saturday’s joint military strike by the US and Israel on Iran, there has been a flood of images and videos purportedly documenting the war. Some are old or reflect unrelated conflicts, were created or AI was tampered withAnd in some cases, is actually derived from Military-themed video games such as war Thunder.
With misinformation spreading like wildfire, many people have placed their trust in reputable digital investigators. organizations like the new York Times, indicatorAnd bellingcat Extensive verification processes are in place to avoid publishing synthetic or misleading content. “Viewers can turn to credible, independent news organizations that take the time and effort to authenticate scenes and clearly explain sourcing,” said Charlie Stadtlander, executive director of media relations and communications. many timestold The Verge. Media authentication methods are rarely foolproof, but the standards are very high, and experts have years of experience in spotting fake news.
This process is no easy task, especially given the lack of reliable deepfake detection tools. But learning from the experts can help us better protect ourselves when news events dominate digital spaces – so here are some of the tricks they use.
Step one: take a very close look
When? unverified images After the US kidnapped Venezuelan leader Nicolas Maduro in January, pictures of him suddenly spread on social media. many times’ visual investigation team jumped into action. They examined the images for visual anomalies “that would suggest they were not authentic” – such as one example that showed a plane with strange-looking windows.
This was certainly not enough to prove that the photos were fake. “But the remote possibility that the photos were not genuine – combined with the fact that they came from unknown sources, and details such as Mr. Maduro’s clothing being different between the two images – were strong enough to disqualify them from publication,” many times’ director of photography Meghan Lurum said in the article.
We’ve largely passed the days of identifying AI-generated deepfakes by counting how many fingers a person has, but there are usually still subtle indicators – for example, check out the architecture and figures in the background. unexplained anomalies.
Step Two: Consider the Source and Its Reputation
An image of Maduro many times The post — showing the Venezuelan leader in custody — came from President Donald Trump’s Truth Social account. This does not mean that Trump or any other government official is a reliable source – they tend to be. The spread of AI fakery The integrity of online, and government, handouts generally can be difficult to establish. Concerns over the authenticity of the image were also flagged, regarding the poor quality of the image and unusually cropped dimensions.
“In this case, the President’s Truth Social post was newsworthy in itself, even though we had no definitive way to confirm that the image was authentic,” Luram said. but it was published many times’ As part of the homepage Screenshot of Trump’s entire postNot in isolation. “Displaying it in context means that, if the image proves inauthentic in any way, we would have presented it not as a legitimate news photo, but as a communication from the President.”
You don’t need to be familiar with the person or organization to recognize potential red flags. An easy way is to check if the account is brand new (or, if it’s old, has no posts before a recent date.) ShowToolsAI and Riddance creator Jeremy Carrasco describes this as “account age paradox”: Because the technology to explain deepfakes is fairly recent, the accounts pushing it were likely created when those AI models were released, and older fakes are easier to spot.
Step Three: Check the Digital Footprint
Sometimes you can quickly debunk fake news by checking if the same photos and videos have been posted elsewhere. You can do this manually by searching for related topics online or using search engine features like Google’s reverse image search tool. The original source may be outdated and completely unrelated to the context with which it is now being shared, as one post claims to show. Missiles attacking Israeli nuclear plant That was actually footage from Ukraine in 2017.
OSINT Platform bellingcat Uses a mix of visual checks, cross-referencing and software tools, including Google and Yandex, for reverse image searches, and extracts metadata from images using ExifTool. However, these investigations are typically time consuming, and are becoming increasingly difficult to sustain due to the increasing accessibility of generic AI tools.
“The flood of fake evidence has escalated things and bad actors have found the ‘it could be AI’ excuse to discredit genuine footage,” said Eliot Higgins, creative director at Bellingcat. The Verge. “Our methods still hold up because we focus not just on pixels, but also on provenance and context, but the noise level is now much higher.”
Step Four: Establish the Date and Location
If a photo or video was reportedly taken at a specific location, you can use satellite images or apps like Google Maps to cross-reference if the location matches. Markers such as flags, logos, and equipment can also be used to determine time periods and locations. many times done in 2022 To verify footage of the Russia-Ukraine conflict. The publications’ investigation team can also guess what time of day the photo was taken Websites like SunCalc Which measures the shadow, and can use footage from nearby CCTV and security cameras to confirm the image.
It is not enough to simply distinguish real photographs from completely synthetic images. How much editing or manipulation is allowed in a photograph before it is considered genuine? No universally accepted answer exists, but Higgins says his personal definition of a photograph is “an actual moment captured by light on a sensor or on film.”
“It’s a testament to what really existed in that time and place. Minor changes like cropping or contrast are fine and always have been, but once you add, remove or create elements (especially with AI), it’s no longer a photo, it’s digital art or propaganda,” says Higgins. “Authenticity resides in honest provenance, not in perfect pixels; that’s why real ground truth images still matter more than any fake.”
“The average person needs to understand that the current information environment is inclined toward manipulation and deception”
Fake news expert and co-founder of open-source intelligence (OSINT) platform indicator Craig Silverman says it’s still important for every online user to remain vigilant. Silverman explained, “The average person needs to understand that the current information environment is inclined toward manipulation and deception. This requires you to scroll with an awareness of how easily images, video, and text can be manipulated.” The Verge. “Add to this the fact that major social platforms have largely failed to deliver on their promises to label AI-generated content, and you get a chaotic, deception-filled, digital landscape that overwhelms and misinforms.”
Everyday people can help stop the spread of misinformation by pausing before sharing anything emotional or viral online. Many of the verification tools that credible newsrooms are using can be accessed by anyone for free. If you don’t want to take action yourself, check any suspicious posts with multiple independent sources.
“Remember that information takes time to develop, especially when it comes to fast-moving conflicts and other news,” says Silverman. “Awareness and patience are key, and they don’t require equipment or expertise. But you have to practice.”
