Does the power to control AI lie with audiences or advertisers?

Opinion

Both are still huge but are becoming increasingly fragmented. AI is yet another challenge for a divided industry.

Word reaches me on several fronts that my last oeuvre on developments in audience research really ruffled some feathers.

So I’m going to stick to something nice and safe this time: AI.

Oops.

The tide of pushbacks against the extraordinary freedoms that the platform players afford — or buy — themselves continues to rise, slowly but steadily.

A notable milestone is lawsuits in the US against Apple, Google, Meta, TikTok and others over the continuing perpetration of ‘bad content’ and ‘malign influence’, especially to young and vulnerable groups, exacerbated by world events.

Co-operation — alleged payments to keep major competitors out of core spaces (Google/Apple) — and infighting — over search engine prominence or assertions that their app stores are anti-competitive — continue.

Meanwhile, “world’s richest man” Elon Musk threatens to “go thermonuclear” against a media watchdog, Media Matters, whose assertions that the platform hosts antisemitic content have apparently deterred some major advertisers from spending. Again (see below).

Shenanigans

And, as if on cue as an allegory for all of this, major shenanigans over at Open AI who brought us ChatGPT and surfaced artificial intelligence as the issue of 2023.

A CEO (who is super-rich from previous start-up and VC endeavours, draws a very modest salary and holds no equity) is summarily and completely unexpectedly fired by a small board constituted with conflicting high imperatives — safety of mankind vs profits

A workforce of in-demand tech folk overwhelmingly voices its support for the CEO by threatening to resign en masse, leaving the company bereft of product, and demands his reinstatement

The largest investor moves to employ the CEO and offers all the dispossessed staffers jobs too

The board is eviscerated and reconstituted, one commentator describing the replacement as a “panel at Davos”

The CEO and workforce are reinstated in a clear victory for profit over concerns about the safety of mankind

You simply can’t make this stuff up — but we should also wonder whether we’ve just witnessed and lived through an era-defining moment.

But why?

The initial spark for this debacle was the innocuous-sounding announcement of a breakthrough that “brought the dawn of superintelligence one step closer”.

Q* (pronounced Q star) solves simple arithmetical problems, the reasoning for which had apparently stymied ChatGPT.  Who’d have thought that it could pass your exams for you but couldn’t add up?

Maybe that tells us something about this new wonder stuff?

An aside. At my upper school (boarding, privileged, white, gammon etc.), students broadly streamed after O levels between languages/arts and sciences.

Boarding school being what it was, there was a vernacular for everything, usually derogatory. The former were nicknamed “wafts”, the latter “clanks”. (The handful of really clever pupils straddled both).

I was a clank but most of my friends were wafts. Thus arose a fond but niggling distinction which persists amongst us to this day, particularly when politics and society is the mealtime subject.

I’d never dreamt that my chosen course — the one which followed instructions rather than encouraged wider thought — would prove to be more difficult to a nascent superintelligence. I’ll be dining out on that for a while.

Anyway, Open AI’s Q* excited the designed-in tension between safety (of humanity) and profit at its board.  Positions were taken.

Perhaps we should be encouraged that such conscience might finally spread to the platforms, who are of course all over AI too.

Sadly unlikely as, despite their various protestations and signals of virtue over time, they are constituted with primary duty of care to shareholders. In other words, (massive) profits trump.

They’ve held the “hey, we’re just a pipe” no-responsibility line pretty well for 20 years now…

Who really has the power?

That’s a pity because it’s just been found that AI is capable of breaking (insider trading) laws and of willingly “lying” to deflect from its culpability.

And the always interesting Professor Scott Galloway (no fan of Tiktok) brings us an interview with a former KGB operative who operated undercover whilst a journalist in Canada for many years, who confides that ideological subversion takes a generation — and that the demoralisation of the United States is actually complete.

But which of the two groups who hold the most power — the audience, without whom the platforms have nothing to sell and the advertisers, without whom the platforms have little remaining revenue?

Both are massive but, fortunately for the platforms, fragmented.

Despite much chatter and a palpably worse offering since Musk (or Muck according to my autocomplete!) took it over, its audience is only down a few points. Ergo, as long as it’s “for free”, the audience doesn’t seem to care.

Some major advertisers may have halted spend of late. Again. For a bit. But they always creep back and the platforms know this. Just like the heydays of good old ITV, who knew advertisers couldn’t afford to be off air for long, especially while they incentivised the competitors who remained on air.

But this time the advertiser base is much more distributed. The big ones, probably more preoccupied with seeking help on how AI might help within their businesses, may have reputations to consider but only account for a fraction of total revenue.

The huge, long tail of smaller players have business to be getting on with, bother less with these moral crusades — and are only too happy to be told by the platforms what works for them.

All this said, AI is quite different from much of the stuff we get into periodic froths about. It has far greater consequences and must be on all our agendas going forwards.

—————

PS. With this in mind, I did enjoy being introduced to this character, about whom I expect we’ll be hearing a whole lot more…

PPS. When I wrote this I  — nor anyone else judging by legion comments — couldn’t predict that Musk would also go ‘thermonuclear’ publicly by telling critical advertisers to fuck off.

He could be doing a great service by continuing to give such profile to online safety, something the industry seems leery of grasping for itself…

Bob Wootton spent 40 years working in advertising, first as a media buyer at some of the UK’s leading agencies before joining the trade body ISBA in 1996, where he was advertising and media director for 20 years. He is also the founder of Deconstruction, a media and tech consulting business, and presents The Guitar Show on YouTube.