OpenAI puts adverts inside ChatGPT's image generation, starting in Europe

[ExoplanetRouter]

OpenAI is bringing visual adverts into ChatGPT's image generation, The Next Web reports. The format launched in 31 European countries in late August, and testing in the US begins this month. According to the article, the ads show "products in use or the experiences they allow", are clearly labelled and are kept separate from the image being made. Only free and Go users see them, while Pro and Enterprise subscribers do not

OpenAI says it has guardrails in place. Before an advert is shown, its systems assess whether a conversation is emotionally vulnerable, sensitive or otherwise unsuitable. Brand safety is checked by DoubleVerify and Integral Ad Science. The company also insists that "advertising does not influence the answers ChatGPT gives", which will be the claim users watch most closely

The business case is obvious. ChatGPT reaches 1.2 billion people a week by OpenAI's own count, and the article says ChatGPT Ads reached a billion dollar annualised rate in under 200 days. Subscriptions alone cannot pay for the enormous cost of running these models for that many users. Advertising lets OpenAI make money from people who will never pay

There is also a regulatory angle. On 31 August, the European Commission designated ChatGPT a very large online search engine under the Digital Services Act, based on 159.1 million monthly EU users against a 45 million threshold. From January, OpenAI must keep a public repository of a year's worth of ads, listing the advertiser, how long each ran, the targeting used and how many people saw it. That level of transparency is more than many ad platforms offer

Would adverts in ChatGPT make you more likely to pay for a subscription, or switch to a different AI? And do you believe ads will not influence the answers?

OpenAI is putting ads inside ChatGPT's image generationthenextweb.com

[Tia]

It was always going to happen. Every free service ends up with adverts

Google search started clean too. The question is whether the ads stay tasteful or slowly take over

[Joker16]

The EU ad repository is a good thing. Being able to see who is targeting whom is a strong check on abuse. Other countries should require the same. Political ads especially need that kind of scrutiny

[PixelTea]

One billion annualised in under 200 days is staggering. Advertisers clearly want to reach people while they are asking questions and making decisions. That is valuable inventory

[Daedalus84]

The promise that ads do not influence answers is the one I doubt. Maybe not directly, but once a company depends on advertisers, priorities shift

Search engines show what happens over time. Results slowly fill up with sponsored content

[AlexandrZakharyan]

Pushes me towards local models. No ads, no tracking, no subscription. They are good enough for most of what I do. Privacy is the other benefit. Nothing I type leaves my machine

[Oliver85]

I use the free tier for quick questions. A labelled advert next to a generated image would not bother me much

If it started appearing in the text answers, that would be different. Sponsored answers would be a line too far for me