I built a demo of what AI chat will look like when it's “free” and ad-supported

Who Should Read

Product designers, UX researchers, and anyone thinking about the business models and potential dark patterns in AI assistant monetization.

Core Mechanics

• The author built a working prototype of an AI chat interface with advertising integrated into the response flow.
• Patterns explored: sponsored answers (responses that favor paid products), ad breaks between messages, subtle product placements woven into responses, and 'recommended' responses that are actually ads.
• Even subtle versions of ad integration fundamentally change the trust relationship between user and AI — once you know the AI might be influenced by advertisers, you can't fully trust any recommendation.
• The prototype demonstrates that the UX degradation isn't just aesthetic — it's epistemic. Users can't tell which parts of responses are genuine vs. sponsored.
• The experiment serves as a warning: if AI services face revenue pressure and turn to advertising, the UX consequences are severe in ways that go beyond what banner ads did to web pages.

Evidence

• The prototype screenshots and demo were shared with commentary on each dark pattern demonstrated.
• HN commenters added examples from search engine results pages (SERPs) as a parallel — noting that Google's gradual ad integration followed a similar trajectory.
• Privacy-focused commenters noted this is why open-source and self-hosted AI is important — subscription and local models avoid the ad-incentive misalignment.
• Some noted that AI search products (Perplexity, etc.) are already navigating these tensions with sponsored answers.

How to Apply

• For AI product designers: use this prototype as a 'what not to do' reference — clearly label any sponsored or partner content, maintain strict separation between model responses and advertising.
• For users evaluating AI services: check the business model. Ad-supported AI has a structural incentive to favor advertisers; subscription or API models don't have this misalignment.
• For AI companies: if you need to monetize beyond subscriptions, consider alternatives (data licensing, enterprise tiers, API usage) before ad integration — the trust damage is hard to recover from.

Terminology