From declaring his hatred of ads to launching them in ChatGPT, Sam Altman has been on quite a journey. But such conversion experiences are common in Silicon Valley—and they offer the best guide to where Altman and his company are about to take us.
As American presidential candidates oppose endless military adventures during the campaign and launch them once elected, so Silicon Valley founders habitually declare their deep distaste for advertising, only to eventually embrace it with the fervour of converts. Just like Trump denouncing forever wars, Bush the candidate railing against nation-building, and Richard Nixon pledging to end the war before expanding it, the world’s most powerful attention merchants were usually against advertising before they were for it.
One of the great ironies of internet history is the story told by Yale researcher Dina Srinivasan about how Facebook prised users away from MySpace in the mid-noughties by promising to respect their privacy and reduce the ad clutter crowding out their screens.
Famously, Larry Page and Sergey Brin disavowed advertising back in their ‘do no evil’ days. “We expect that advertising funded search engines will be inherently biased towards the advertisers and away from the needs of the consumers,” the duo wrote. That was 1998. In 2025, Alphabet raked in an extraordinary $312 billion in advertising revenue.
Amazon was already the world’s biggest and most profitable supermarket when Jeff Bezos decided to turn his customers into a product. The man who once said “advertising is the price you pay for having an unremarkable product or service” now owns the world’s fastest-growing ad platform, with a remarkable US$69 billion in advertising revenue in 2025.
So maybe the best indication that Sam Altman was about to create an advertising platform was that he kept saying how much he couldn’t stand advertising. After all, this is what the candidates always say.
But there was an ardour in the OpenAI founder’s public statements that appeared hard to walk back. ‘I will disclose, just as a personal bias, that I hate ads,’ he confided to an audience at Harvard University in 2024. ‘I think that ads-plus-AI is sort of uniquely unsettling to me’.
Even as the massive infrastructure costs underlying the chatbot mounted and he was pressed about how OpenAI stops losing billions each year and starts to meet a staggering $1.4 trillion in reported compute costs, Altman remained dismissive. ‘We have a very simple business model and I like it, and I know that I’m not the product… I like that people pay for ChatGPT and know that the answers they’re getting are not influenced by advertisers,’ he maintained.
So even allowing for the precedent set by his predecessors, the speed of Altman’s transformation over the last twelve months has been a sight to behold. In March 2025 he openly mused that ‘maybe there’s a tasteful way we can do ads’; then an internal report surfaced the following month, forecasting revenue from “free user monetization” surging to a whopping US $25 billion by 2029.
By Christmas, OpenAI was playing whack-a-mole, strenuously denying reported sightings of ads on ChatGPT as they surfaced around the internet. Finally, in January, the company posted an announcement that it would start testing ads. And now the ‘test’ is clearly over. OpenAI is teaming up with anybody who can bring in the world’s biggest advertising clients, along with the audience targeting capacity and the measurement and attribution models that they demand.
Having acknowledged that the combination of AI and advertising seems deeply disconcerting and openly contemplated the ‘dystopic visions of the future’ that could unfold, Altman has quite a bit of explaining to do. He can hardly have been surprised when rival Anthropic skewered the move with a Super Bowl ad featuring a personal trainer who starts slipping paid promos into his pep talk. Altman had already publicly set out the ad’s premise years before, railing against a scenario ‘where you ask ChatGPT something and it says, ‘Oh, you should think about buying this product’’.
Altman insists that his company has set out a series of principles that will keep its reasoning model and ad business separate. The guidelines will supposedly ensure that: ads are always separate from the chat and clearly labelled; ads won’t influence the answers ChatGPT generates, and that OpenAI ‘will not optimize for time spent in ChatGPT’ but instead ‘prioritize user trust and user experience over revenue’.
They are nice principles. But when a company starts selling eyeballs only a matter of months after its founder was discoursing on how much he hates advertising, it would seem you can only put so much stock in them. It is difficult to read these latest commitments without an asterisk and a very short use-by date appearing in the mind’s eye.
The deeper point here is about much more than Sam Altman. Just as American presidents find themselves carried along by the imperatives of empire, even against their own better judgment, the forces that drive the attention economy are larger than individual actors. In significant respects, the story of the internet is a tale of how new technological possibilities have been shaped to meet the demands of its wealthiest and most powerful customers—the marketers who pay top dollar for our attention. This imperative determines which companies survive and thrive, and which perish; and—as Altman’s embarrassingly hasty metamorphosis illustrates—how they capture and commodify our attention.
As the inexorable process of enshittification has progressively engulfed each platform, advertising has served as a powerful laxative. The imperative to deliver the right kind of user attention at the lowest cost has trumped every other consideration, even when it conflicts with the interests and experiences of the digital public. Over time, anything that gets in the way of funnelling our attention towards ads, mostly served up by the world’s most powerful corporations, is gradually downsized, diminished and degraded.
The best guide to where OpenAI is headed is not its stated principles but the platforms whose founders once, like Altman, rhetorically resisted advertising revenue only to eventually adopt it as their raison d’être. Where Google started out with a mission to ‘organise the world’s information’, today, when users search terms like ‘How to reduce greenhouse gas emissions?’, ‘eco-friendly companies’, and ‘net zero’, one in five of the adjacent ads are from big fossil fuel firms. Over time, the label that distinguishes an ad from a search result has gradually been diluted to the point it has become barely noticeable, and many digital denizens do not recognise the difference between actual organic results and ads from Big Oil.
Infamously, Instagram’s algorithm is so heavily geared to maximising time spent scrolling that, according to Meta’s own internal research, simple queries for ‘healthy eating’ and ‘healthy recipe’ can lead to pro-eating disorder and even pro-self-harm content within weeks.
On Amazon, the pursuit of the ad dollar means that the top search result is now typically 29 percent more expensive than the best deal, and the best deal is an average 17 places down, underneath all the paid results.
There is evidence that OpenAI is already on a similar trajectory, a process that is best described as the Metafication of ChatGPT. The first step was the recruitment of former top Meta marketing executives, Kate Rouche and Fidji Sumo, to head up the company’s monetisation operation, with 630 former Meta employees (of OpenAI’s 3,000) following in tow.
If bringing in the big hitters from Facebook’s glory days of polarisation and misinformation was less than reassuring, next Altman began pointing to Instagram as a shining example of how to do digital advertising. ‘I love Instagram ads,’ he enthused in October. ‘I found stuff I never would have found. I bought a bunch of stuff.’
The valorisation of Instagram makes zero sense if the goal is to ‘prioritise user trust and experience over revenue’. No other platform has such a clear record of cannibalising its own users in the pursuit of profit. On the other hand, as a signal that OpenAI intends to reorient its entire platform around the requirements of advertisers, Instagram works perfectly.
ChatGPT is already experimenting with a range of products that could replicate Instagram’s presently unrivalled power to funnel attention all the way from ‘brand discovery’ to the buy button. It appears that the rudiments of Altman’s ‘dystopic vision’, in which marketers use AI to insidiously insert themselves into the most intimate and personal parts of our lives, are falling into place.
In the stories we tell about the internet, the trade in audience attention is usually treated as a backdrop, like a feature of the natural world. In the foreground, heroic founders unleash the forces of disruption, or technology itself is personified, turned into a character and made to bear our sins. The reality is that new tools and the people who invent them come to reflect the underlying economic imperatives. If we are to assert social control over the internet, we need to interrogate the end to which it is largely devoted, the ceaseless commodification of attention.