AI phobia is the new American consensus 

Financial Times Europe19 Aug 2026Edward Luce edward.luce@ft.com

Stop us before we end human ity,” the AI titans keep telling us. Also, “trust me andme alone, not my com pet it ors”. Whether it is Meta’s Mark Zuck er berg, OpenAI’sSam Alt man, xAI’s Elon Musk or even Anthropic’s Dario Amodei, Amer ica’s techbro l ig archs are their own worst alarm ists. In the same breath as warn ing thathumans are head ing for the evol u tion ary dust bin, they vow to end dis ease, abol ishwork and fix global warm ing. Call it death cult uto pi an ism. Amer ic ans are nowincreas ingly focused on the death cult part.

Little the tech chief tains say should be taken at face value. Every utter ance must befiltered through the lens of mar ket rivalry. Put simply, the AI com pan ies that areper form ing best — not ably Anthropic, which is pre par ing for AI’s largest ever ini -tial pub lic offer ing — are in favour of reg u la tion and closed-weight mod els, or thosewith private para met ers. They want a Food and Drug Admin is tra tion for AI.

The lag gards, on the other hand, have found reli gion in open-weight mod els. Hav -ing spent his life striv ing for Face book mono poly, Zuck er berg is now warn ing aboutthe dangers of mar ket con cen tra tion. But he suf fers from low pub lic trust even bySil icon Val ley’s plum met ing stand ards. Young Amer ican women refer to Meta’s AIRay-Bans as “per vert glasses”.

Barely a week passes without one of the bro l ig archs pub lish ing an AI mani festothat dresses their cor por ate imper at ive in human ity’s lar ger interests. Tellingpeople that your com pany’s tech no logy is par tic u larly dan ger ous is another way ofsay ing that your stock will out per form the oth ers. Plus, “I am more far-sighted thanmy rivals about the risks”. Little won der that Amer ic ans no longer trust their techcham pi ons.

No soci ety on Earth has his tor ic ally been more pro-tech no logy than Amer ica. Theworld should thus take note of its Lud dite turn. Almost three quar ters of Amer ic ansdo not trust busi nesses to use AI respons ibly, accord ing to Gal lup. More than 70 percent oppose hav ing data centres built in their area because of fears about waterand elec tri city infla tion. Eighty per cent or more dis trust AI for driv ing, med icaladvice and cor por ate hir ing decisions. As The New York Times recently put it, fearof AI is the “most bipar tisan issue since beer”.

That feel ing is as strong in the centre as it is on the Maga right and the demo craticsocial ist left. That such fears are often con spir at orial should be no sur prise. Para -noia is a nat ural response to the unknown. Maga’s Steve Ban non calls data centres“weapons labs”. Mar jorie Taylor Greene, the former pro-Trump law maker, refers toBig Tech as “Skynet” after the self-aware com puter sys tem that trig gers nuc learholo caust in the Ter min ator movies. Pro gress ives talk of AI killing US demo cracyand ush er ing in a Blade Run ner-style dysto pia.

Since the AI titans are almost to a man addicts of sci ence fic tion, they can hardlycom plain that its tropes are being used against them. Larry Page, Google’s formerchief exec ut ive, reportedly called Musk a “spe cies ist” for wor ry ing that AI super in -tel li gence would leave human ity in the dust. Every AI fig ure, at some point, hasspec u lated about immor tal ity or mul ti plied life expect ancy. I, for one, do not rel ishthe idea of hav ing to listen to them after I turn 150.

A key to mar ket ing is to per suade con sumers that they need something of whichthey were pre vi ously ignor ant.

Polit ics, on the other hand, is about giv ing voters what they say they want. Bothrules of thumb have been thrown off kil ter by AI. Con sumers are being enlightenedabout something they dread — loss of human autonomy. As voters, however, theirfears are being ignored. US polit ics is there fore enter ing a rad ic ally unpre dict ablephase.

Pres id ent Don ald Trump is unchar ac ter ist ic ally deaf to voter fears on this. Much ofhis hos til ity to AI reg u la tion is about money. Musk spent hun dreds of mil lions onTrump’s 2024 cam paign and is prom ising another $100mn or so for Repub lic ans inthis year’s midterm elec tions. Most of the AI com pan ies have con trib uted to vari ousTrump boon doggles, includ ing his White House ball room and his Free dom 250 cel -eb ra tions of Amer ica’s anniversary.

Trump is mean while warn ing the world that you are with the US or against it. Forcoun tries such as Singa pore, the United Arab Emir ates and, increas ingly, Europe asa whole, such ulti mat ums are unwel come. They want the free dom to bene fit fromboth China and Amer ica’s com pet ing and var ied tech no lo gical plat forms. Each ofthe super powers is build ing its own global AI alli ance. Any attempt to reg u late AI inAmer ica is dis- missed as help ing China.

Here are three things a decent US admin is tra tion of either party would do. First, itwould respond to soci ety’s legit im ate fears by set ting up a pub lic-private part ner -ship on AI safety. It would take a leaf from the Pope’s recent encyc lical on put tinghuman dig nity first. Second, it would shift the US tax bur den from labour to cap ital.Cur rent US fiscal incent ives strongly favour the automa- tion of jobs. Third, it wouldtry to negoti- ate AI prin ciples with China to stop the world from a bin ary deathspiral.

These are bare min imum steps. Our spe cies still has scope to shape its own future,this fic tional US pres id ent would say — but that win dow is clos ing fast.

Does this sound unreal istic? Alas, at the moment it does. Now, on the other hand, isa golden moment for US politi- cal entre pren eurs