The article on WIRED discusses the concept of “enshittification,” a term coined by writer and tech critic Cory Doctorow, which describes the process by which companies initially build platforms that serve users effectively, but degrade in quality once they achieve market dominance in order to maximize profits. This phenomenon has been observed in major tech platforms such as Google, Amazon, Facebook, and TikTok. Doctorow’s ideas gained wide attention when his essay on the subject was republished by WIRED, leading to “enshittification” being named the American Dialect Society’s 2023 Word of the Year. Doctorow further explores this phenomenon in his book titled after the term.
The article reflects on the potential for artificial intelligence (AI) to fall into the same trap of enshittification. AI, particularly models like GPT-5, are becoming integral in various aspects of daily life, including providing recommendations for dining, interpreting events, and making purchasing decisions. The concern is that as AI becomes more embedded in everyday activities and the costs of developing these technologies soar, the few companies dominating this advanced sector might prioritize recovering their massive investments over serving user needs. This could lead to AI recommendations being influenced by paid promotions rather than user interest—an advertising model some AI companies are already considering.
OpenAI, an influential player in AI development, recently partnered with Walmart to integrate ChatGPT into the shopping experience, raising concerns about conflicts of interest. While some AI companies, such as Perplexity, assure that ads will not compromise their commitment to unbiased service, the potential for such boundaries to erode remains a concern. OpenAI CEO Sam Altman expressed hopes to develop advertising products that are beneficial to both users and the company, yet the skepticism remains high given the financial incentives.
Beyond advertising, Doctorow points out other ways in which platforms have shifted their business models once they secure a dominant market position. He cites the example of Unity, a leading videogame development tool provider, which introduced a new “runtime fee” that was met with significant backlash and eventually retracted due to user revolt. Similarly, services like Amazon Prime Video introduced ads and continually increased subscription costs, further exemplifying how successful platforms might exploit their user base to extract more value.
Doctorow argues that AI, particularly because of its opaque, “black box” nature, possesses an even greater potential for enshittification. He suggests that the intensive capital demands of developing AI might push companies to degrade their service quality or introduce new costs for users sooner rather than delivering initial value. This possibility is alarming considering AI’s deep integration into personal and professional spheres, where it holds significant influence over information consumption and decision-making processes.
Interestingly, when the author queried GPT-5 about the potential for AI to undergo enshittification, the AI itself acknowledged that Doctorow’s framework fits well with AI systems if unchecked incentives persist. This response from GPT-5 underscores the inherent recognition of potential issues within AI systems themselves, suggesting that without careful regulation and ethical considerations, AI might indeed follow the path of other tech platforms that prioritize profit over user benefit.
In conclusion, while AI technologies like GPT-5 currently provide significant value and convenience to users, there is a palpable concern that these benefits could diminish as companies seek to recover hefty investments and maximize profits. The debate around AI’s future frequently returns to the need for vigilant oversight and ethical practices to prevent the degradation of service quality—an enshittification—that could betray user trust and diminish the technology’s potential positive impact on society.
Read the full post on wired.com
