There is a basic dignity owed to a person after they are killed, and a large part of it is simple: you show people who they actually were. A national news network failed that test in January 2026, not through malice but through the increasingly common act of putting an image on air without checking whether it was real.

What happened

Alex Pretti was an intensive-care nurse with the Department of Veterans Affairs. He was killed by border agents in Minneapolis, during a period of heavy immigration enforcement in the city that was already generating a flood of contested images and video. On January 26, 2026, MS Now, the network formerly known as MSNBC, used a photo of Pretti during a segment of Deadline: White House hosted by Nicolle Wallace.

The photo was not a real photograph of him. Someone had taken a low-quality image, reportedly sourced from the VA, and run it through a generative AI model, apparently to produce a clearer, higher-resolution picture. What the model produced was not a faithful enhancement; it was an invention. It broadened his shoulders, tanned his skin, and reshaped his nose. The cumulative effect, as critics put it bluntly, was to make a dead man look more conventionally handsome than he was. The network aired this fabricated likeness on its broadcast, its website, and its YouTube channel. The Daily Mail and International Business Times ran the same altered image.

"We didn't make it" is the problem, not the alibi

When Snopes asked about it, MS Now's parent company, Versant Media Group, said the network had obtained the image from the internet without knowing someone had altered it, and that it had not enhanced the photo itself. That is almost certainly true. It is also the entire failure stated as if it were a defense.

A newsroom's job, the part that distinguishes it from a group chat forwarding pictures, is verification. "We found it online and did not realize it was fake" is a precise description of the thing professional journalism is supposed to prevent. The network did not need to be the one wielding the AI tool to be responsible for what it broadcast. It put an unverified, AI-altered image of a real, named, recently killed person on a national news program, and the only reason anyone knows is that viewers and outside observers noticed something was off about the face.

This is what separates an incident like this from someone deliberately fabricating an image to deceive. No one is claiming MS Now set out to mislead. The harm came from a verification process that has not kept pace with an information environment now thick with AI-altered material. The model did exactly what these tools do, generate plausible pixels, and the institutional safeguard that was supposed to catch it, a human asking "is this actually him," did not run.

The backlash and the quiet swap

Criticism arrived quickly and from across the spectrum, including from podcaster Joe Rogan, who called out the network for beautifying Pretti. The objection was not partisan; it was about the basic wrongness of broadcasting a cosmetically AI-altered version of a man who had just been killed, as if his real face were not good enough for television.

MS Now responded by replacing the video's thumbnail and updating the description with an editor's note acknowledging that the previous thumbnail had featured an AI-enhanced image of Pretti. That is the appropriate correction, and it is also a small monument to how late the check came. The note exists because the image aired first and was scrutinized second, which is the reverse of how this is supposed to work.

The bigger environment this lives in

Pretti's altered photo did not happen in a vacuum. The Minneapolis enforcement surge produced a wave of AI-altered photos and videos that blurred the line between real and fabricated, including a separately circulated AI image that falsely depicted Pretti holding a gun. Newsrooms covering fast-moving, emotionally charged events are now swimming in a supply of imagery that has been quietly run through generative tools by people with every range of motive, from "I wanted a sharper picture" to deliberate disinformation. The old reflex, grab the clearest available image and run it, is now a liability, because the clearest available image is increasingly the one that has been synthetically smoothed, sharpened, or invented.

The lesson, stated plainly

This belongs on a list of AI failures on two counts. A model asked to sharpen a low-resolution photo did not sharpen it; it invented a different, more flattering face, which is wrong output dressed up as an upgrade. And professional outlets then let that AI-altered likeness reach the public as fact, because the editorial verification that was supposed to catch it did not run. The fix is not technological and not new. Before a face goes on air attached to a real person's name, somebody has to confirm it is that person's actual face, sourced from something traceable, not lifted from an internet copy that may have passed through a generative model on its way to you. In an environment where any image might be synthetic, "we got it online" stops being an explanation and becomes a confession. The dignity Pretti was owed was small and specific: to be shown as himself. A national network, and the outlets that followed it, could not clear that bar, because nobody checked before they aired.