Can AI avoid the trap of obfuscation?

I recently went on vacation In Italy. As one does these days, I ran my itinerary past GPT-5 to get sightseeing suggestions and restaurant recommendations. The robot reported that the best option for dinner near our hotel in Rome was a short walk down Via Margutta. It turned out to be one of the best meals I can remember. When I got home, I asked the model how she chose this restaurant, which I hesitate to reveal here in case I want to get a table sometime in the future (hell, who knows if I’ll even come back: it’s called Babette’s. Call ahead to reserve.) The answer was complex and impressive. Among the factors were rave reviews from locals, notices in food blogs and the Italian press, and the restaurant’s famous blend of Roman and contemporary cooking. Oh, and the short distance.

There was something also required of me: confidence. I had to buy into the idea that GPT-5 was an honest broker, choosing my restaurant without bias; That the restaurant didn’t show me as sponsored content and didn’t get a portion of my check. I could have done a deep dive on my own to double-check the recommendation (I already searched the site), but the point of using AI is to get around that friction.

This experience strengthened my confidence in the results of AI but also made me wonder: As companies like OpenAI gain more power, and as they try to pay off their investors, will AI be vulnerable to the erosion of value that seems endemic in the technology applications we use today?

Play on words

Technology writer and critic Cory Doctorow calls this erosion “cluttering.” His premise is that platforms like Google, Amazon, Facebook, and TikTok start out with the goal of pleasing users, but once companies outcompete competitors, they intentionally become less useful in order to reap greater profits. After WIRED magazine republished Doctorow’s groundbreaking article in 2022 on the phenomenon, the term entered the vernacular, mainly because people realized it was completely true. Enshittification was selected as the 2023 Word of the Year by the American Dialect Association. This concept has been cited so often that it goes beyond profanity, appearing in places that would normally turn up noses at hearing such a word. Doctorow has just published an eponymous book on the subject; The cover photo is an emoji of…guess what.

If chatbots and AI agents become angry, it could be worse than Google search becoming less useful, Amazon results being plagued by ads, and even Facebook showing less social content in favor of anger-generating clicks.

Artificial Intelligence is on its way to being a constant companion, providing quick answers to many of our requests. People already rely on it to help interpret current events and get advice on all kinds of purchasing choices, and even life choices. Given the enormous costs of creating a complete AI model, it is fair to assume that only a few companies will dominate this field. They each plan to spend hundreds of billions of dollars over the next few years to improve their models and get them into the hands of as many people as possible. For now, I would say that AI is in what Doctorow calls the “good for users” phase. But the pressures to recover huge capital investments will be enormous – especially for companies with a constrained user base. These conditions, Doctorow writes, allow companies to abuse their users and business customers “to extract all the value for themselves.”

When one imagines the beauty of artificial intelligence, the first thing that comes to mind is advertising. The nightmare is that AI models will make recommendations based on which companies have paid for the hire. This is not happening now, but AI companies are actively exploring the advertising space. In a recent interview, Sam Altman, CEO of OpenAI, said: “I think there are potentially some great advertising products that we can do that are a net gain for the user and kind of positive for our relationship with the user.” Meanwhile, OpenAI just announced a deal with Walmart so that the retailer’s customers can shop within the ChatGPT app. I can’t imagine the struggle there! AI search platform Perplexity has a program where sponsored results appear in clearly labeled follow-ups. But it promises, “These announcements will not change our commitment to maintaining a reliable service that provides you with direct, unbiased answers to your questions.”