I recently took a trip to Italy and sought recommendations from GPT-5 for sightseeing and dining options. One suggestion, a restaurant in Rome named Babette, turned out to be exceptional, showcasing a delightful blend of Roman and modern cuisine. However, while I enjoyed this experience, I found myself questioning the trustworthiness of AI recommendations. Was GPT-5 devoid of bias, or was the recommendation influenced by third-party interests?
As companies like OpenAI seek to monetize their powerful AI models, I pondered whether these tools might suffer from a process described by author Cory Doctorow as "enshittification." Doctorow’s concept states that tech platforms begin with the goal of serving users but shift their focus to profit maximization after eliminating competition, ultimately degrading user value. His theory gained traction after WIRED published his 2022 essay, leading to its recognition as the American Dialect Society’s 2023 Word of the Year. Doctorow further expanded on this idea in his recent book, aptly titled "Enshittification."
If AI tools were to follow the same trajectory as other tech platforms, their decline in quality could have severe implications. As AI becomes a staple in everyday decision-making—providing insights on current affairs and consumer choices—there’s a growing concern that profit pressures will lead to compromised integrity. Doctorow’s term captures this fear: companies will prioritize their bottom line over user satisfaction, extracting maximum value for themselves.
The potential for enshittification in AI is alarming, particularly with the looming prospect of advertising infiltrating AI recommendations. OpenAI’s CEO, Sam Altman, has suggested that monetizing AI through advertisements could be beneficial, sparking fears of biased recommendations based on financial incentives. Furthermore, AI firms are already experimenting with sponsored content, raising questions about maintaining trust and objectivity.
Doctorow warned that once a company possesses the capability to compromise its product’s integrity, the temptation to do so becomes irresistible. He pointed to examples like Unity, which attempted to implement a controversial new fee structure, and how streaming services have increasingly integrated ads into previously ad-free experiences.
While Doctorow remains skeptical about AI’s current value, he acknowledges that enshittification may occur even in early developmental stages. The complexities of AI can allow companies to obfuscate their degradation of service, potentially leading to significant user abuse.
Engaging with GPT-5 led to a surprising alignment with Doctorow’s concerns, as it noted that the framework of enshittification applies disturbingly well to AI systems if unchecked financial motives prevail.
In summary, while I’m grateful for AI’s assistance in enhancing my travel experience, the broader implications of potential decline in AI quality due to profit-driven motives raises significant concerns for the future of these technologies. If history proves anything, it’s that we must remain vigilant to resist the allure of enshittification before it seeps into our daily lives.
