Five reasons why ads in LLMs are no big deal

I read a bunch of posts last week that were shocked by the progress of ads in ChatGPT and other chatbots. Why is everyone so surprised? Didn’t they think the industry would be quick to lean into ads in large language models (LLMs)? 

The fact is, even if the ads are not overtly present, their influence is. In my opinion, it’s better to embrace ad integration than to pretend it doesn’t already exist.

Rather than engage in shock and outrage at the speed of adoption, let’s look at the facts. Off the top of my head, I can think of five key reasons why ads in LLMs are actually a very good thing.

The ad industry thrives on innovation

Most of the channels at our disposal are quite mature, and the rapid growth of artificial intelligence (AI) has hurt search more than any other channel, while video continues to grow.

We could use a new channel to work with, and this is the single fastest-growing form of media any of us has ever seen. Of course we must work with it.  Of course we are going to find ways to put ads into it.

The line between content and ads is already blurry

Much of the growth in advertising and marketing has been in transforming “content” into a promotional vehicle, especially in social media and video.

These are formats used to feed the LLMs, so a significant portion of the “content” analysed and referenced in their responses is indeed advertising content. It is promotional in nature.

The LLMs may get better at sorting through viable third-party content vs. what a brand or a paid party has created, but the fact is, it’s difficult to sort through it. They are already influencing the responses, for better or worse.

LLM responses are already being aggressively influenced

Remember SEO — search engine optimisation? It has quickly expanded into GEO, or generative engine optimisation.

They are already working the systems and trying to get brands mentioned more here or there, so why not be up front about it? Let’s accept that every new channel becomes the target of influence, and simply find ways to be clear about the integration of commercial efforts.

This is an inevitable way to start making money

The AI chatbots were always going to be monetised with ads. If you thought otherwise, you were fooling yourself.

Only so many people will ever pay for a subscription. There were always going to be ads integrated in some way. I personally wish they would follow the display model and integrate visually, but I also realise how much of the UI is now voice-activated, and that voice model needs monetisation, too.

The faster we embrace a model, the sooner consumers can become smarter

The other fact is that consumers understand advertising. If you get them used to ads in the user interface (UI) now, while day-to-day adoption is increasing, they will be used to it.

They can also figure out when advertising is in the mix. They will become better at distinguishing what is and isn’t an ad and respond accordingly.

Advertising in this channel was always an inevitability. I don’t care what the powers that be were saying. It was always going to be the end game, and it’s better to start toying with it now — letting brands and agencies be part of the solution, as well as end users.

That way, we can come to a resolution that works in the long run and will drive ROAS for people with the money to support it.