Times Insider explains who we are and what we do and delivers behind-the-scenes insights into how our journalism comes together.
The image was stunning, newsworthy — and in need of a judgment call.
Hours after President Trump announced on social media Saturday morning that the United States had captured Nicolás Maduro, the leader of Venezuela, he posted an image that appeared to show Mr. Maduro in sweatpants, blindfolded and handcuffed onboard a U.S. warship.
It was Mr. Trump’s attempt to prove a mission accomplished. “Nicolas Maduro on board the USS Iwo Jima,” Mr. Trump captioned the image. But Mr. Trump had a history of sharing A.I.-generated images on social media. My team of photo editors had to assess the image he posted and help make the call of whether to publish it.
Photo editors in Seoul had been working on the Venezuela story since news of explosions there broke. They had been coordinating with colleagues in London and New York on the deployment of photographers in Caracas; plans to station photographers in New York, where we expected Mr. Maduro to arrive; coverage of reaction among Venezuelan communities and potential protests; and Mr. Trump’s scheduled news conference.
What we had not planned for was the sudden proliferation of various online images of Mr. Maduro.
Before Mr. Trump’s posted the photo, an editor alerted us to a different image circulating on social media that appeared to show Mr. Maduro detained by American troops or officers from the Drug Enforcement Administration. The image was unverified, provided by a source of a New York Times reporter. Clinton Cargill, the news director for our photography department, found that image, and another, on social media.
At that point, The Times and other news outlets were reporting that Venezuela’s vice president was calling for “proof of life” of Mr. Maduro, an additional factor in assessing the news value of any images of the ousted leader.
Clinton used an A.I.-detection tool and consulted with Stuart Thompson, who writes about disinformation. Both flagged those two images as having inconsistencies that would suggest they were not authentic. For instance, Stuart noted an odd-looking second row of windows in the aircraft hold.
Still, several A.I.-detection sites did not find definitive signatures that the images were generated by A.I. These sites are not foolproof, however; they often register some uncertainty with images that are verifiably authentic.
But even the remote chance that the images were not genuine — coupled with the fact they came from unknown sources, and details like Mr. Maduro’s clothing being different between the two images — was strong enough to disqualify them from publication.
While Clinton was investigating these photos, Mr. Trump posted the image of Mr. Maduro.
It looked odd: The photo was cropped to an unusual, vertical shape, suggesting that much of the content of the original image had been excluded, and the quality was low. One Times photo editor noted that it looked like a photo of a printout, or a photo of a screen. When run through the same A.I. detector, the image posted by Mr. Trump also left some uncertainty.
The Times has reported on Mr. Trump’s habit of disseminating A.I.-generated imagery and deepfakes on social media, so we had reason to be skeptical of the authenticity of the photo.
We are also cautious about publishing government handout images, as we cannot necessarily verify their authenticity, nor can we vouch for the journalistic rigor with which they were made, even if they are genuine. And we are mindful that images distributed by any government are often intended to advance a specific narrative.
On the rare occasions that we do publish government handouts, it is because the image or its distribution is highly newsworthy, or because a photo was captured in a place we do not have access to. We always clearly label these images as handouts and describe their provenance in a caption.
In this case, the president’s Truth Social post itself was newsworthy, even if we had no surefire way to confirm that the image was authentic. Julie Bloom, the head of our Live news desk, was interested in publishing the image along with a blog item about the president’s post, and our editors in charge of the print newspaper wanted to publish it in the Sunday edition.
We made the judgment call. We decided that the best way to present the image would be to show it in the context of Mr. Trump’s Truth Social post, rather than isolating the image. Displaying it in context means that, if the image proves to be inauthentic in some way, we will not have presented it as a legitimate news photo, but rather as a communication from the president.
We decided to publish a cropped version of the post on The Times’s home page. In print, we published the full post on an inside page. The permanence of the print front page made us careful about featuring it prominently.
The authenticity and credibility of our news report is always paramount, and the tools for detection are vital for our work. Still, currently there is no tool that unequivocally verifies images. Like so much in journalism, it is up to us — human editors — to make judgment calls and to provide our readers with information they need to know, with the appropriate context and caveats.
The post How The Times Assessed That Photo From Trump of Maduro in Handcuffs appeared first on New York Times.
