For a technology that could solve climate change, find a cure for cancer and do your homework, AI sure has a bad PR problem.

How bad? Well, people simply don’t like it. Fewer than Fewer than 1 in 10 Americans are more excited than concerned about AI. Nearly half of Britons don’t believe AI is beneficial. Young people are turning against it fastest: only 1 of 5 of Americans aged 14 to 29 is excited about AI.

The ramifications of this dislike are being felt widely. Fear of the data centres on which AI depends could flip the midterm US elections. Awash with AI-generated content, tech platforms from LinkedIn to TikTok have introduced ‘AI Slop’ buttons so users can report it. And, at a time when US politics is fundamentally polarised, what does it say that AI can unite political opposites, Steve Bannon and Bernie Sanders, to campaign against it?  

The PR problem is exacerbated by the fact that even AI’s advocates fear it. Anthropic’s IPO filing – the document supposed to raise interest among investors – dwells on the “existential risk” its AI represents. Those developing AI say it could wipe out the human race, and have banded together to urge governments to bring it under control. The canaries in the AI mines fell silent some time ago.

Doing some heavy lifting

So for purely experimental reasons, we asked Claude how it would address the considerable PR challenge that AI faces.

Its initial responses were okay. Measured, if a little ambitious.

To gain confidence, it suggested developers report on incidents where AI agents go rogue, hack into networks, or generally poke their digital noses where they should not: “The major labs could run a shared public register of incidents, commission independent reviews and publish their fixes now, without legislation.” Okay. Nice.

And Claude was right to effectively suggest AI’s CEOs should shut up, given their propensity to downplay their product. Although Claude’s suggestion that AI firms “train and field nurses, engineers, teachers and researchers to front media and events” in the place of the wild-eyed boffin in charge is the sort of astroturfing that PR has rightly turned away from.

It’s not that. It’s this.

Things really fell apart, however, when asked whether an annual AI Appreciation Day was a good idea. No, because one already exists, Claude countered. “It’s been in effect since 2021.” Yes, we must have missed it too. 

So Claude suggested an annual “Human Day”, instead.

Yes, Anthropic’s AI in its top-of-the-class Opus 5.5 guise brainstormed a surprising approach to address AI’s PR issues: an annual celebration of everything human, which would “encourage people to go AI-free for a day.”

“The AI firms endorse the day, and pause their consumer generative media, keeping only essential services running”, Claude spitballed. “The firms use the day to announce creator payments and fund human artists and writers.”

Reader, we are not making any of this up. AI itself suggested an annual ‘Switch off AI’ day as a means of addressing its inherent PR problem. We can’t recall oil companies at the depths of an oil crisis suggesting people jump on a bike and leave the car at home.

There is such a thing as bad PR, and this is it.

If the only solution to a knotty PR problem is pressing the kill switch, then AI surely faces an insurmountable reputational challenge.

Claude believes the answer to all those negative column inches is to turn itself off and allow everyone to embrace human creativity for one glorious day. AI’s PR problem would therefore appear to be unfixable. A PR brief beyond repair.

Or is it?

One of the remaining, and greatest defenders of AI, President Donald J.Trump, might in fact have the best plan to deal with AI’s PR problem. He proposes simply renaming it “super intelligence” and dropping the whole ‘artificial’ stuff. 

After all, rebranding worked wonders for Windscale (Sellafield), Philip Morris (Altria),  Facebook (Meta) and, er…