Artificial intelligence is in its “good to users” phase — powerful, accessible, and often free.
But history shows that platforms rarely stay that way. As Cory Doctorow warns, once a technology gains dominance, it often enters a cycle of enshittification — prioritizing profit over users until the product degrades.

This post outlines ten early warning signs that AI systems may be following the same path.

⚠️ 10 Warning Signs of AI Enshittification

1. Sponsored or Biased Answers Begin to Creep In

When AI responses start favoring paid partners or “recommended” brands without clear disclosure.

Example: An assistant suggests one travel site over another because it’s a paid partner.

2. Tiered Access to “Real” Intelligence

Free users get throttled models while premium users retain quality.

Creates information inequality and pressures users into paywalls.

3. Increased Friction for Switching or Exporting Data

AI platforms make it hard to export chat history, fine-tunes, or projects.

Lock-in comes first; extraction follows.

4. Loss of Transparency or Auditability

Less visibility into data sources, model versions, or influences behind answers.

Opacity enables hidden monetization and bias.

5. Hidden or Expanding Surveillance

Platforms quietly collect more user data “to improve performance.”

Shifts from serving users to monetizing them.

6. Degradation of the Free Tier

Free versions become slower, inaccurate, or stripped of features — not due to cost, but to drive upgrades.

Echoes social media’s shift from user reach to pay-to-play.

7. Monetization Over User Benefit

Product design choices prioritize ads, upsells, or partnerships over usefulness.

Sponsored content in AI output is a classic symptom.

User data, conversations, or uploads are used for model training or shared with partners without explicit opt-in.

The user becomes an unpaid contributor.

9. Declining Accuracy or Usefulness Over Time

As models become optimized for corporate interests, responses lose clarity and trustworthiness.

Utility quietly erodes.

10. Community and Developer Ecosystem Suppressed

APIs close, plug-ins vanish, and open-source competition is throttled.

Innovation and user autonomy fade as ecosystems are walled off.

🧭 How to Spot the Shift Early

  • Compare answers across multiple AI systems regularly.
  • Watch for pricing and tier changes.
  • Look for “sponsored” or “partnered” language in responses.
  • Read Terms of Service updates carefully — words like personalization or improvement often mean expanded data use.
  • Track the subtle downgrades: slower speed, less detail, fewer export options.

📚 References

This article is part of an ongoing series on technology ethics and AI design integrity.
Originally inspired by WIRED’s coverage and Doctorow’s framework.