Everybody hates AI

I’m preparing for what must be close to my 20th trip to Cannes Lions later this month. I’ll be publishing a Cannes Diary from the event. It won’t be dry panel recaps or odes to marketing bravery, I promise. The Rebooting is holding several activations in Cannes. Our lineup:

If you’re interested in invitations to these activations, please register your interest here.  Hope to see you there.

ChatGPT’s using "heartwarming retro vibes” to convince people that AI is not something to be feared but a trust companion to improve everyday life. New ads show it being used to impress a date with lemon-garlic butter pasta. Another shows a wholesome brother-sister road trip – with an itinerary planned by ChatGPT. This is very cautious marketing from an industry that’s been telling us its building a digital god. All this capex for recipes and itineraries? What happened to curing cancer?

OpenAI and the entire AI industry has a major problem on its hands. The massive amounts of capital they’ve poured into AI have not led to miraculous breakthroughs. Everywhere warning signs are flashing. AI is polling worse than ICE. Young people hate AI because they see it compressing their future prospects. Workers hate AI because its makers say it will take their jobs. Parents hate AI for screwing up their kids’ education. Journalists hate AI Companies hate AI pricing because they’re not seeing productivity gains for all that spending. Managers hate the tsunami of AI workslop that’s invaded corporate life. Even LinkedIn hates AI on LinkedIn. Creative people hate AI for commoditizing art. The pope warns against AI entrenching power and furthering exclusion at the expense of human flourishing. 

Narrative matters more than ever, and the AI narrative has shifted. Silicon Valley has bungled the rollout of this “discovery.” It obsessed over weirdo sci-fi narratives around AGI rather than focus on practical examples of AI improving human flourishing. It marketed the “promise” of AI replacing doctors and teachers. Tech oligarchs were its spokesmen in a populist time. And it has dismissed opposition to data centers as part of a Chinese subversion. No wonder everybody hates AI.

The resistance is likely futile in the long run – there’s too much money and power to be accumulated – but the pressure building from various corners of society will have an impact on the diffusion of AI. This is how societies work: messily. The AI YOLO era is over.

David Sacks was able to water down the AI safety regime but not stop it. The White House buried the announcement. Hardly on-brand for an administration that makes a spectacle of everything. They were dragged into this by mistaking AI as a tech issue. It is a general purpose technology that has impact across many, if not most, parts of society. That means it intersects with many constituencies and power centers. Anthropic’s new Mythos model was a wakeup call. The tentative regulatory regime is the proverbial camel’s nose in the tent.

A besuited Sam Altman is meeting Bernie Sanders to discuss AI concerns. Sanders wants the government to take 50% of the  shares in AI companies to use for the benefit of the American people. A compute tax funneled back to citizens might be a more workable model. Altman needs to burnish his image ahead of the early 2027 release of Artificial, a Hollywood treatment of his battle for AI that will portray him as a duplicitous scoundrel. We’ll also be treated this fall to “The Social Reckoning,” a sequel to “The Social Network” that will dwell on the adverse consequences of the social media era.

All this together is a new direction in the AI narrative and the narrative around tech’s role in society overall. At its best, technology has encouraged human flourishing even as it has caused disruption and displacement. We are in an extended and belated trough of disappointment over the entire nature of consumer AI as a result of the social media and mobile era being widely seen as a societal disaster. The scorecard is actually far more mixed, but the issue of social + phones on kids is enough to give normal people pause about handing too much power to the very same people. 

Regulation is now a reality. Kevin Hassett made the ultimate Washington gaffe by accidentally speaking the truth when he said AI would have an FDA-style regime. That is inevitable. The first baby steps were taken with a White House highly averse to any regulation imposing a mandated 30-day review of new models. The tech industry is already entwined with the government across many vectors. 

The only question is timing. Regulate too soon, you’ll squelch competition and lock in winners in a government-protected oligopoly. I found it telling that the All-In crew has zeroed in on antitrust as regulation has become an inevitability, even somehow claiming Anthropic is a monopoly. Their discussion of Magnifica Humanitas skipped almost entirely the substance to instead zero in on the specter of regulatory capture. 

The media industry has been in a defensive posture with AI. There have been lawsuits, of course, notably The New York Times vs OpenAI, but for the most part, publishers are involved in a negotiation. There’s a business cost to being anti-AI. Instead, there’s been grumbling about how unfair it all is. It is now safe to call a spade a shovel.

The New York Times publisher Arthur Sulzberger this week warned of tech oligarchs of “hijacking the public square” with “brazen intellectual property theft… at an unprecedented scale.” "Our profession has been too quiet, too passive and too fragmented in the face of abuses by the companies leading the A.I. revolution," he warned. That’s more muscular. In past comments about AI, Sulzberger has been more measured, going out of his way to laud AI’s “great potential” and argue for fairness through some kind of regulatory regime. This week marched a shift in tone.

The tricky part is balancing that kind of messaging while internally telling employees they have to master these tools. The New York Times is battling with its unions over the use of AI.  McClatchy execs are warning of ruin if they don’t use AI to drive higher output more efficiently. This can be seen as anti-AI sentiment in a hidebound profession. But it’s also the street knowledge of journalists who saw the Boss Class for a generation respond to digital disruption with a call for MORE CONTENT. It never worked, hard to see how using a Content Scaling Agent™ is the answer now. The answer, as Sulzberger noted and The New York Times has shown, is better journalism. 

Thanks for reading. Send me a note with your feedback by hitting reply.