Automating the mass shooting victim

On AI as a false simulacrum of hope for change. Plus: A copyedit chief on being 'replaced' by AI, Elon Musk is voted least popular newsmaker in America, and the Luddites in National Geographic.

Jim Acosta’s “interview” of an AI-generated mass shooting victim is, if nothing else, the ideal artifact for the days of the peak AI boom. Send this thing to the Smithsonian, put it on loop in an exhibition hall next to the Ghiblified Twitter avatars and the AI-generated witness testimony that led a judge to hand down a harsher sentence.

To back up: Acosta, CNN’s former White House correspondent and Trump irritant, who’s recently launched a Substack and YouTube show, aired a segment in which he questions an AI-generated simulacrum of Joaquin Oliver, one of the teenagers killed in the 2018 Parkland high school massacre. So on the occasion of Joaquin’s would-be 25th birthday, Acosta asked his chatbot reproduction about gun violence, Star Wars, and how it would address the mass shooting epidemic.

It’s not just that “former broadcaster interviews an AI-generated school shooting victim” is a construct bleaker than any cyberpunk story beat I can think of offhand, it’s how many trend lines towards decline had to intersect to have made this particular event possible. Critics have piled on, and not just because it’s another callous ‘reanimate the dead with subpar AI’ stunt. There’s a lot packed into the hull of this one, and it’s so much more tragedy than farce.

Brief obligatory note that Blood in the Machine is a 100% reader-supported publication. Subscribe below. Supporters who part with $6 a month make my work possible—big thanks to all of you—and also get access to what I’ve been calling the Critical AI Report below the fold. Today, we’ve got Elon vs the AI critic pope in the polls, a brand new AI scam, Luddites (and yours truly) in National Geographic, and the chief copyeditor of Business Insider speaks out on being “replaced” by AI. Thanks all, and onwards.

---

First, who knows if we ever would have seen this at all if Acosta hadn’t gotten demoted and left CNN. Speculation abounds, but the network was reportedly planning on moving him to a midnight time slot, in a move some saw as a gesture of appeasement to Trump. If Acosta was still at CNN, the network might have balked at airing something like this for legal, ethical, or simple commercial reasons. There is certainly a media criticism story here, about how the battered standards of professional journalism are making room for dubious stuff like this.

As for the AI itself, it’s, let’s say, very bad. AI-generated Joaquin’s movements are stilted and crude, its voice well out of sync. That voice is monotone one minute and lurching into a pitch shift the next, delivering canned lines in a sentence structure anyone who’s used a chatbot for a few minutes will recognize. It’s an animatronic and poorly photoshopped ghost, and the effect might be comic if it weren’t brought into being by such profound tragedy.

You have to sympathize with the parents. They’ve endured a devastation I almost called unimaginable but has become imaginable to too many parents who’ve lost children this way. The Olivers have been stalwart campaigners for gun reform, and they created the AI avatar of their son out of a desire, they say, to continue to raise Joaquin’s voice. They want to bring his AI-generated avatar onstage to policy debates, to start social media accounts for him, to make him unignorable, and also so they can continue to relate to their lost son. It’s all so unspeakably sad.

Critics of Acosta’s AI interview have called this “grotesque” and so forth but it’s easy enough to understand the desire. An American politics in thrall to the gun lobby and warrior capitalism has proved incapable of confronting the epidemic of mass death, and no matter how many children are gunned down, grieving parents form political advocacy groups to seek reform, or former mass shooting survivors attempt to enter politics only to see the establishment eject them, nothing at all changes.

To these grieving parents, maybe this AI offers at least a gesture towards a horizon where some kind of change, anything at all, is possible. Even if that promise is illusory, and the bot quite technically incapable of offering any new insights, trained as it is on corpuses of data drawn from the same gun wracked past. It is some kind of motion, forward or not.

Who knows whether the parents, after fighting seven years for common sense gun reforms and getting nowhere, even realistically think this AI will make a difference. But it’s something, even if it’s a numb, disconcerting shrug of a something. Come to think of it, while we’re wallowing in the bleakness here, a numb, disconcerting shrug of a something is a serviceable descriptor for the appeal of AI as a vehicle for social improvement.

When tech industry lights say “AI will solve climate change” or “AI will cure cancer,” prospects that also only sound feasible or even aspirational in the cold light of broad institutional failure, does anyone believe them anymore? Does anyone take that pitch at all seriously? More likely they are giving off a numb disconcerting shrug, because though AI might not even be a credible panacea, at least it’s enough to generate the illusion that progress is possible. For plenty of investors, that’s probably good enough.

Most of the rest of us, I think, intuit that the automated ghosts of mass shooting victims won’t offer any durable insights that might help us get out of these messes. If anything, by appropriating the remaining political will to technological fantasia and further tethering us to products sold by companies on an antidemocratic streak of their own—they’ll dig us further in.

The Critical AI Report 8/5/2025

Elon Musk is the least popular “newsmaker” in America, and AI critic pope is #1

Pretty amusing stuff from a new Gallup poll here that finds that the most popular man in America is our new tech critic pope. Guess which two figures come in dead last?