In 2024, Sam Altman called ads in AI a dystopian idea.
He painted the picture: you ask your AI assistant something, and instead of helping, it nudges you towards a product. He said he preferred subscriptions specifically to avoid users becoming the product.
Now OpenAI is running ads. It was always on the roadmap. The pressure to monetise was inevitable and the moment has arrived.
Yesterday's post on this sparked some really insightful comments. A lot of concern around how these ads could be manipulative given how much trust people place in ChatGPT.
Once people know answers are biased by advertising, I’d like to think they'll start second-guessing everything. Look at how differently people perceive Apple vs Google. One is built for consumers, the other for advertisers. That gap exists for a reason.
I'm not saying all of this will be bad. Some of these ads will be genuinely useful, the most personalised we've ever seen. There are Instagram ads right now that surface things people actually want. That same thing will happen inside ChatGPT.
But there's a fine line between helpful and manipulative. And if history is any guide, we won't stay on the right side of it for long.
Some ideas came up on how to get this right:
Contextual over personalised. Let people opt out of deep audience targeting. Only use the context of that specific chat, not everything they've ever asked.
Quality scores on ads. Rank ads on usefulness and value, not just who pays most. Something like Clearcast in the UK, where ads need approval before they run. Only ads that genuinely add value make it through.
Brand building over manipulation. Treat it like TV; a leaned-in, high-attention environment for brand building. Not lower-funnel nudges designed to exploit trust.
Ads aren't inherently bad. They democratise access to information and keep products free.
But when advertiser incentives outweigh consumer needs, the product stops serving you. It starts serving whoever's paying. Money talks. It always has.
The dystopia Altman warned about isn't some far-off future. It's arriving now. And the question isn't whether it will happen, but how far it will go before we push back.
I posted a video of him talking about this. Funny how quickly things change when the runway gets shorter.
Quoting a post byCameron Armstrong ↗