How long have we got before AI agent outputs are enshittified?

A digital illustration metaphorically depicting the future of AI agents and the platform degradation process
AI Summary

Examines the causes and prospects of 'enshittification,' where AI agents and platforms secure monopoly power using convenience as a weapon, only to degrade into ads and user data collection.

Imagine waking up one morning, turning on your smartphone, and having your AI agent (an artificial intelligence assistant that performs complex tasks on your instructions) effortlessly handle your schedule and personalized recommendations. As if by magic, it selects only the answers that suit your taste. The world is all about these convenient AI tools, and we spend our days chatting with this agent.

However, this convenience leaves a strange sense of unease. Thoughts like, “Will this great service be free forever?” or “Will someday all my conversations and preferences be plastered with billboards?” Serious discussions on this question are already taking place in the tech industry [Link to reddit.com]. How long do we have before our AI agents break down and start producing ‘enshittified’ outputs?


Why does this matter?

Countless digital services we use initially offer tremendous benefits to users. They are free or radically cheap, ad-free, and seem to care only about our convenience. However, once time passes and the service dominates the market, things change completely.

This phenomenon goes beyond simply disliking a single app. At a point where our daily lives, work, and even ways of thinking have come to depend on AI agents, if the service degrades, users become trapped in a massive prison from which they cannot escape even if they want to. The “privacy nightmare” where platforms suck up all user information and sell us off to advertisers becomes reality [Will My Tech Enshittify? How to tell… - The Opt Out Project]. Therefore, understanding how this cycle of change works is an essential defense for modern digital citizens.


Easy to Understand: What is Enshittification?

Writer Cory Doctorow, who analyzed this phenomenon most sharply, explains it with the word “Enshittification” [Enshittification - Wikipedia]. Put simply, it is “the process whereby services progressively become a mess.”

To use an analogy, imagine a very friendly supermarket opens in your neighborhood, giving away items for free. Everyone flocks to that supermarket. Existing neighborhood stores all go bankrupt and close. The moment that supermarket becomes the only shop in town, the owner’s attitude changes. Prices rise, quality drops, and the store becomes filled with advertising flyers. But because there are no other alternatives, residents have no choice but to go there reluctantly.

In economics, this is described as a platform acting as an intermediary while simultaneously exercising a monopoly over suppliers and a monopsony over consumers [Enshittification - Wikipedia]. Even when substitutes are visible, the state where the entire ecosystem is locked in and one cannot go elsewhere is called switching costs [Enshittification - Wikipedia]. The world of AI agents is walking down the exact same path.


Current Situation: The Tug-of-War Between AI Agents and Platforms

Currently, voices of concern regarding the future of AI agents are erupting among tech communities and developers worldwide. Tech journalists like Rach Smith constantly raise the question: “While we are intoxicated by the sweet outputs of AI agents, when will this system degrade into a tool for advertising and monetization?” [How long have we got before the agent output is enshittified?].

In online communities like Reddit, discussions are also heating up: “What kind of advertising model will be attached after AI is fully popularized and occupies a monopolistic position?” and “Is our privacy safe?” [Link to reddit.com].

Many experts warn that when tech products and services transition from early growth stages to monetization stages, the pattern of collecting information by mortgaging user trust and pushing targeted ads repeats itself [Will My Tech Enshittify? How to tell… - The Opt Out Project]. The cold reality is that the smart AI agents we use today are unlikely to escape this grand law of capital.


What Will Happen Next?

Then how should we cope amidst this massive wave of degradation? While we cannot predict the future completely, we must be able to detect certain clear signals in advance.

  1. Acceleration of Data Collection: Be wary if the agent’s feature of remembering your preferences becomes excessively detailed. There is a high probability that all your private data is being absorbed into corporate servers in the name of convenience [Will My Tech Enshittify? How to tell… - The Opt Out Project].
  2. Loss of Neutrality: It will become difficult to distinguish whether the restaurants, products, and information recommended by AI agents are truly useful to you or filtered by algorithms of companies that paid hefty advertising fees.
  3. Finding Alternatives and Preparing to Escape: Because high switching costs make it hard to escape once locked in, maintaining ongoing interest in open-source AI models or decentralized services is important.

MindTickleBytes’ AI Reporter Perspective

Technological advancement always presents us with new freedom and convenience. However, the receipt for that gift is always billed later. Whether AI agents remain our reliable companions or turn into shrewd middlemen targeting our wallets and information ultimately depends on the critical eye and choices of us who use the technology. It is a time when we need an attitude of constantly questioning and being wary of what the price of convenience is.


References

  1. How long have we got before the agent output is enshittified?
  2. Enshittification - Wikipedia
  3. Pluralistic: How twiddling enshittifies your brain
  4. Will My Tech Enshittify? How to tell… - The Opt Out Project
  5. Link to reddit.com
AD
Test Your Understanding
Q1. What is cited as the main reason why platforms act as intermediaries and form monopolies and monopsonies?
  • High switching costs
  • Government deregulation
  • Spread of open-source AI
Even if substitutes exist, users and services cannot leave due to high switching costs, forming monopolies and monopsonies.
Q2. What is the concept explained by Cory Doctorow called, where a platform degrades by collecting user information and selling ads?
  • Enshittification
  • Open-sourcing
  • Decentralization
Enshittification refers to the process where a platform initially focuses on users, gradually lowers intermediary value, and degrades into advertising and data collection.
Q3. Which of the following is NOT a common side effect when tech products or services degrade?
  • Excessive information collection
  • Selling advertisements
  • Enhanced user privacy protection and free security
Degraded services cause a privacy nightmare and focus heavily on data collection and ad sales.
How long have we got before...
0:00