Mark Zuckerberg's 'AI is great, really' PR campaign is in full swing. Here's a better idea.

• Mark Zuckerberg is doing a big AI PR push.
• That includes a new ad, op-eds and interviews for old media outlets, and a Wall Street pep talk.
• It's unclear if any of that will move the needle. Why not try cold, hard cash?

Many Americans are very concerned about AI. Can Mark Zuckerberg change their minds?

That is what the Meta CEO is trying to do, via a persuasion campaign: Last week, the company launched an ad, scored by a David Bowie song, telling viewers to ignore "some people" who don't like AI.

Zuckerberg followed up a Wall Street Journal op-ed on Tuesday, promising that "superintelligence will be the most profound technological advance we will see in our lifetimes," and that this will be a good thing. Then he got on the phone with reporters at the Journal, the Financial Times, and The New York Times to say the same thing.

And on Wednesday, Zuckerberg used his company's earnings call to keep the pitch going: His company's AI efforts make him "optimistic that we're building a positive future for everyone," he told investors.

There are at least two things going on here:

• Zuckerberg, like lots of leaders in AI, seems taken aback by the AI backlash in the US. Like many of them, he thinks the industry needs a better communications plan. One that sells the upside of AI, and downplays the not-so-good stuff.
• Zuckerberg, like lots of people in AI, seems particularly frustrated with OpenAI and Anthropic, whose CEOs have warned repeatedly that AI will bring about severe dislocation, including meaningful job losses. There's a school of thought that OpenAI's Sam Altman and Anthropic's Dario Amodei have overstated AI's risks in order to help justify their enormous fundraising and valuation ambitions. So when Zuckerberg writes "I don't understand why anyone who believes that AI will eliminate most jobs and much of humanity's relevance would rush to build that future," in his WSJ op-ed, he's talking about Altman and Amodei.

Let's leave aside whether Zuckerberg, or anyone at Meta, is the best advocate for the AI industry (in early 2025, two-thirds of Americans reported having a negative view of Zuckerberg).

My bigger question is whether anyone can convince Americans to feel more positive about AI.

It is certainly possible that Americans would feel better about AI if Altman and Amodei hadn't spent time warning them that the tech they were building could generate massive job losses. They might also have felt better if the broader tech industry, which is now essentially the AI industry, hadn't publicly pledged allegiance to Donald Trump starting in late 2024.

But those things happened, and I don't think a messaging campaign is going to make people forget them. At this point, the way to convince people that AI is going to be good for them is to demonstrate that good things are actually happening.

That's because right now, most of the AI upside is either theoretical or intensely concentrated among a few people — primarily people who work or invest in AI. If you're selling a home in San Francisco, it's easy to see an AI benefit. For just about everyone else, it's something that might help you someday.

So while the industry waits for other upsides — New wonder drugs! New industries! Things so new we don't have words to describe them yet! — to materialize, my advice to Meta, and everyone else who wants to change public opinion about AI, is simple: Pay up.

Teachers in rural Louisiana, where Meta is building a massive data center, are getting one-time bonuses of up to $50,000 this year. That payment doesn't come directly through Meta, but via a huge tax payment Meta made as part of its buildout.

$50,000 doesn't mean much to a Meta engineer. In Richland Parish, where Meta is building its data center, that money could be a year's salary or much more.

So Mark, if you want people to feel good about AI, and what you're doing with AI: Do more of that.

Let all of the smart people who work for you figure out the best way to do that: Direct payments to citizens? Money that goes to local government first, and then gets turned into personal checks? Or maybe you just build a bunch of schools and roads and make sure everyone knows who paid for them.

Whatever it is, do it fast — you're good at fast — so people know that AI is helping them now, not one day in the future. It might work better than videos and op-eds.