Don’t let Advertising Ruin AI
AI experiences like DALL-E, ChatGPT, and more are the most exciting new technologies in nearly two decades. There are several reasons why these advances are so exciting, but perhaps the main one is that they reveal that search engines are being “hired” by users to solve problems. However, the economics of scale demand that we ask the question: who will pay for these experiences? The answer could undermine what makes AI most useful.
An anamorphic film frame of the film “AI Mad Men (2032)” depicting a robot being attacked by Google, 8k — Ryan & DALL·E
Google has eroded the value of the internet
In my career, I have worked at several companies that have needed to devote a lot of attention and resources to placating Google in order to keep their businesses running. Google knows this well, since almost everything it does is subsidized by the money generated by its search advertising. To keep their business growing, Google has mastered the art of keeping search engine optimization (SEO) a moving target that requires constant study and monitoring.
The only way many companies can keep up is by employing specific SEO experts. These professionals are embedded on teams that serve as de facto Google ambassadors. Google Search’s market share practiallydemands these technologists ignore most non-Google search engines and is responsible for companies continuing to chase Google’s algorithms, for fear that not being found will spell the end of their website traffic.
To make matters worse, Google has eroded the value of the internet with its snake oil-like algorithms that entire industries follow. This leads to bloated content and hidden answers. In a desperate search to be found by Google’s users, website cater to vague rules that determine “content quality” often resulting in an experience that leaves people hunting through search results and webpage content stuffed with Google-appeasing filler. You can think of this as the “internet recipe effect.”
Here’s the thing…
People never actually needed search results; people want their problems solved — and AI experiences are just better set up to do that.
AI is too tempting for advertisers
I predict that it will be almost impossible to avoid giving in to one of the most basic and popular models of monetization — advertising — when it comes to widely available AI experiences.
Advertising ruins user experiences. The core business of advertising is at odds with honesty and objectivity — someone can pay money to intercept a moment for the user and leverage that moment to sell something they might not otherwise know about or purchase.
Advertisers will flock to AI-powered experiences that have large audiences, willing to pay to have their products/services be the “default” answers and recommendations to users looking for answers to problems.
“How should I repair a crack in my wall?”
“What are the best shoes for someone with wide feet?”
“Show me some options for great notebooks for daily planning and research.”
If powerful companies can pay to change answers — and questions like these come with the asterisk of “well, this isn’t the REAL answer according to people, it’s actually what a brand paid for you to see” — the answers become invalid.
If an AI cannot provide a user with an answer to a question without being manipulated by brands willing to pay, we will have hobbled the potential for our next great human advancement before it even has a chance to transform our lives for the better. Imagine Google, Amazon, Facebook, or any search experience that doesn’t have results that are manipulated in some way by advertising. They would all instantly be better.
Everything comes down to cost
The resources needed to power today’s AI models currently has plenty of costs associated with developing, maintaining, and running them, and recent trends have indicated that costs aren’t going down as significantly or quickly as they have in the past. Removing all profit from the pipeline — there is just raw electricity, time, and labor cost to develop, maintain, produce, and serve technology.
Divorced from capitalism, technology is at its most useful when as many people can access it as possible. The problem for technologists always comes down to “How do I get this to as many people as possible?” Thanks to how most societies operate, technology always has more people who need it or benefit from it than are willing or able to pay for it.
We’ve already seen how “Web 2” handled the issues of cost. Today’s internet is largely powered by free products and services powered by advertising. I think we must find a better way forward to realize a future that is actually better for the people who will use this new technology.
To everyone who is actively working to bring these things to the world: good luck, don’t fuck this up.
