In both literal and figurative sense. And it will happen because companies needs to make money.
There are things about dynamics on the internet that we have taken for granted. And we assume them to be true whilst they no longer are true.
One of those is that we assume search engines give us the relevant content. It never was the case, but it was good enough for us to get incredible usefulness to be able to find useful content, but it wasn’t the most relevant. No ranking model is perfect, but they were much better than none and guessing websites.
Sides were mostly happy about it, as they were giving information in exchange for click troughs, where they could monetise you. And we have seen this breaking with search engines showing summaries of a journalists works, limiting click troughs and not necessary paying for the sources. We start to see breaking of the soft contract of why that model worked. Then they have added promoted content and ads that meant that the ranking model started to shift towards firms that could pay more for advertisement.*
At this point what have not been straight forward is to pass that advertisement cost directly onto you. And that is about to change.
What has been your experience of internet in this era? You have searched, went to major website or found a niche forum on page 2-3 of search results. You got feeling that it is showing everything that internet has to offer. And for most of us, that has been the case. Then it started to feel a bit odder. Multiple ads, websites you expected to be there weren’t showing, results that looked like ads, so you were looking at five ads, five results whilst only one felt sensible and then five more ads. It stopped being the door to the internet, whilst our brain hasn’t really picked up the change of incentives behind the scenes.
Sides can exclude themselves from search results. From being scraped. And we have seen it in a past. Such as big news channel arguing with the search engines as they wanted they share of revenue.
Even trough all of this, what search engines have given us was a sense of context. You saw multiple results, you saw ads, you saw a lot of the results were weird. It still feels like you have some agency and make judgement. You can intuitively go to the next page and see if something useful is there or use different search engine with hope of better results.
You saw a bit further than only your necessary result and you could have follow curiosity and you would have better understanding of what your possibilities are. It gave you context to judge the quality of the result. It is the same when you are browsing to buy things at your favourite online shop. You get promoted products that the company is getting paid to promote, but you will see others. It gives you something to investigate and change your mind.
And we have seen with the social media the trend to limit what you see. Being serve one photo and video at the time with limited control of what is displayed. All of us are in some content bubble because of it. We don’t have any wider context to make the judgement that we are in the bubble, because you don’t see it.
The lack of wider context is what makes it hard to judge the quality of completeness of the information. Are you getting balance content, political views, products based on quality? Likely not.
AI Agents are another move in this trend. Limit the information available. Which is a bit ironic when LLMs are training on all the available data. And you can argue that AI gives you a lot of relevant answers with links to the sources, so what am I talking about?
Now, let’s go back to the original premise why this exists. Who is making money on this? Those are not government run services like healthcare in Europe. Companies have to pay for a lot of expensive infrastructure and people and shareholders want return on investment. Capitalism 101.
The change will be driven by the break of the contract that originally search engines made with the content providers. You give us data and we will get you people to your website. Now, with AI, these sides are not getting click troughs, they get much more bot traffic and possibly might pay to show ads to bots. So when you don’t have people traffic from AI agents that you could monetise, you have to charge for the content provided and for it to be available for AI agent to scrape and summarise.
Companies not wanting their sides to be accessible for AI agents is already happening and I think it will become a common situation.
Tell me, do you think, when you ask your AI agent a question that it browses trough ALL the relevant sites? What if the most relevant content is the most expensive to produce and that revenue stream was cut off? What if, instead of AI Agent giving you the relevant content written by a journalist, it will give you summary of Reddit comments under and article (biased based on subreddit where it poped up)? Notice how often you see Reddit in the answers and one of the reason is because it licenses its data.
The challenge is that we will never know what data is and isn’t licensed and accessible to the AI Agent. It could become possible that yIn both the literal and figurative sense. And it will happen because companies need to make money.
There are things about dynamics on the internet that we have taken for granted. And we assume them to be true whilst they no longer are true.
One of those is that we assume search engines give us the relevant content. It never was the case, but it was good enough for us to get incredible usefulness to be able to find useful content, but it wasn’t the most relevant. No ranking model is perfect, but they were much better than none and guessing websites.
Sites were mostly happy about it, as they were giving information in exchange for click-throughs, where they could monetise you. And we have seen this breaking with search engines showing summaries of a journalist’s work, limiting click-throughs and not necessarily paying for the sources. We start to see the breaking of the soft contract of why that model worked. Then they added promoted content and ads, which meant that the ranking model started to shift towards firms that could pay more for advertisement.*
At this point, what has not been straightforward is to pass that advertisement cost directly onto you. And that is about to change.
What has been your experience of the internet in this era? You have searched, went to a major website or found a niche forum on page 2-3 of search results. You got the feeling that it is showing everything that the internet has to offer. And for most of us, that has been the case. Then it started to feel a bit odder. Multiple ads, websites you expected to be there weren’t showing, results that looked like ads, so you were looking at five ads, five results, whilst only one felt sensible, and then five more ads. It stopped being the door to the internet, whilst our brains haven’t really picked up the change of incentives behind the scenes.
Sites can exclude themselves from search results. From being scraped. And we have seen it in the past. Such as big news channels arguing with the search engines as they wanted their share of revenue.
Even through all of this, what search engines gave us was a sense of context. You saw multiple results, you saw ads, you saw that a lot of the results were weird. It still feels like you have some agency and can make a judgement. You can intuitively go to the next page and see if something useful is there, or use a different search engine with the hope of better results.
You saw a bit further than only your necessary result, and you could have followed your curiosity and you would have a better understanding of what your possibilities are. It gave you context to judge the quality of the result. It is the same when you are browsing to buy things at your favourite online shop. You get promoted products that the company is getting paid to promote, but you will see others. It gives you something to investigate and change your mind.
And we have seen with social media the trend to limit what you see. Being served one photo or video at a time with limited control of what is displayed. All of us are in some content bubble because of it. We don’t have any wider context to make the judgement that we are in the bubble, because you don’t see it.
The lack of wider context is what makes it hard to judge the quality or completeness of the information. Are you getting balanced content, political views, products based on quality? Likely not.
AI agents are another move in this trend. Limiting the information available. Which is a bit ironic when LLMs are training on all the available data. And you can argue that AI gives you a lot of relevant answers with links to the sources, so what am I talking about?
Now, let’s go back to the original premise of why this exists. Who is making money on this? These are not government-run services like healthcare in Europe. Companies have to pay for a lot of expensive infrastructure and people, and shareholders want a return on investment. Capitalism 101.
The change will be driven by the break of the contract that search engines originally made with the content providers. You give us data and we will get you people to your website. Now, with AI, these sites are not getting click-throughs, they get much more bot traffic and possibly might pay to show ads to bots. So when you don’t have people traffic from AI agents that you could monetise, you have to charge for the content provided and for it to be available for an AI agent to scrape and summarise.
Companies not wanting their sites to be accessible to AI agents is already happening, and I think it will become a common situation.
Tell me: do you think that when you ask your AI agent a question, it browses through ALL the relevant sites? What if the most relevant content is the most expensive to produce and that revenue stream was cut off? What if, instead of an AI agent giving you the relevant content written by a journalist, it will give you a summary of Reddit comments under an article (biased based on the subreddit where it popped up)? Notice how often you see Reddit in the answers, and one of the reasons is that it licenses its data.
The challenge is that we will never know what data is and isn’t licensed and accessible to the AI agent. It could become possible that you will want it to go through regulatory documentation on a government website, but that site will block any AI agent’s access, so it will just generate it from its training dataset and link the site.
Whilst in the search you could see 15 results next to each other, with an AI agent you might see one or two, but they will give you much higher confidence that they are the ones you are looking for. By nature of the user interface and the language used.
Imagine how much companies will pay to be the ONE AND ONLY result shown to you. Imagine how much companies will pay to be the ONE AND ONLY running sneaker shown to you and immediately bought. With all the text to convince you that it is the best option for you.
Imagine how much more they will be able to charge you. How much will you pay for a sneaker if the AI agent tells you that it’s the best option and the best price, whilst they were paid to be shown to you this way? Will you go out of your way to do a good ol’ search around multiple websites to compare it? Or will you pay an extra 10% for the convenience to just click and buy at that moment?
We will see dynamic pricing become a regular occurrence. Likely each of us will see a different price, or we will all see a rise in cost to cover these technologies and costs.
Shopping through an AI agent will become them auctioning you to find out how much you are willing to pay. And the way it will happen is that you won’t see the context. You won’t be able to compare unless you go out of your way. And we’ve seen on modern social media that it is not easy to go out of your way.
How do you know that the restaurant that will be booked is the one you might actually enjoy, or the one that can pay the most to the AI agent to get you to book it?
How do you know that the holiday the AI has selected isn’t the one with the most empty spaces they need to fill, and you will pay extra for it as well? The only way of knowing would be using different services and browsing around for you to understand that your result might not be objective.
The only way for you to know if something is a “fair deal” is to understand the context of the deal. Past price, reviews, similar products and services, without that information being skewed by the monetary incentive.
The ability to limit the context of information shown by AI agents and the incentive for companies to pay for their services to be shown will be more powerful than ever, and it will come with a price tag that we will pay for.
*You might see a lot of scam and fraudulent ads, because their limit of how much they can get from you is your whole bank account. Rather than journals that might charge you £5 a month. Fraudsters are big spenders.ou will want it to go trough regulatory documentation on government website, but that will block any AI agents access, so it will just generate it from it’s training dataset and link the site.
Whilst in the search you could see 15 results next to each other, with an AI agent you might see one or two, but they will give you much higher confidence that they are the ones you are looking. By nature of the user interface and the language used.
Imagine how much companies will pay to be the ONE AND ONLY result shown to you. Imagine how much companies will pay to be the ONE AND ONLY running sneaker shown to you and immediately bought. With all the text to convince you that it is the best option for you.
Imagine how much more they will be able to charge you. How much will you pay for a sneaker if the AI Agent will tell you that it’s the best option and best price, whilst they were paid to be shown to you this way? Will you go out of your way to do good ol’ search around multiple websites to compare it? Or will you pay extra 10% for the convenience to just click and buy at that moment?
We will see dynamic pricing to be a common thing. Likely each of us will see different price or we will all see rise in a cost to cover these technologies and costs.
Shopping trough AI Agent will become them auctioning you to find out, how much you are willing to pay. And the way how it will happen is that you won’t see the context. You won’t be able to compare unless you go out of your way. And we’ve seen on modern social media that it is not easy to go out of your way.
How do you know that the restaurant that will be booked is the one you might actually enjoy or the one that can pay the most to the AI agent to get you to book it?
How do you know that the holidays AI has selected isn’t the one with the most empty spaces they need to fill and you will pay extra for it as well? The only way of knowing would be using different services and browse around for you to understand that your result might not be objective.
The only way for you to know if something is a “fair deal” is to understand the context of the deal. Past price, reviews, similar product and services without that information being skewed by the monetary incentive.
The ability to limit the context of information shown by AI agents and the incentive for companies to pay for their services to be shown will be more powerful than ever and it will come with the price tag that we will pay for.
*You might see a lot of scam and fraudulent ads, because their limit of how much they can get from you is your whole bank account. Rather than journals that might charge you £5 a month. Fraudsters are big spenders.
