There's a black mirror episode like this. Not quite as bad having an AI doing it, though.

A woman needs this treatment after a brain injury. Iirc it clones her brain onto their servers. It's super experimental, it's a free surgery with a subscription service where you need to be within the supported area like cell towers.

If she stops paying for the service she will die. They promised to expand the support area to more countries and areas, but they end up raising prices as well as adding in advertising.

The advertising takes over her body at random contextual times and she basically does an ad read to the people in whatever situation.

It starts happening at the school where she works, she tells a kid about a certain product after he asks her about a problem. She's gonna lose her job because it creeps everyone out, so she has to pay more to get rid of the advertising.

https://en.wikipedia.org/wiki/Common_People_(Black_Mirror)

I'm not sure it would be so easy to do it in a consistent, verifiable way. You can certainly prompt the LLM to work the ad into the conversation, but making sure it actually happens and is done in a way that the advertiser is going to be happy with is a lot harder.

> It would be so easy, to generate subconcious addds with an LLM. Like weave a slogan reminder into the conversation...

Probably not without changing the underlying response and tanking the benchmark ratings. Like I ask it to generate code, it tries to add a slogan reminder, now the comments on my code have Tide's slogan in them and the variables are all named "bleach".

Probably less humorous than that, but the underlying point stands. People will be extremely upset about their prompts being tinkered with, and any perceived degradation in quality will be quickly attributed to them. The hidden nature of them will probably cause speculation that they do the same on the higher tier plans.

When ethics become hilarious.

Now I'm wondering if all those extra Ds in ads are a subconscious ad.