Advertise in ChatGPT
OpenAI’s move to introduce contextual ads in the free tier of ChatGPT is being read as both a predictable monetization step and a worrying sign about the company’s financial pressures. Commenters fear a familiar “enshittification” path like Google and YouTube: ads that start clearly labeled and separate from answers but eventually blur into or bias the model’s output, undermining trust. Others see this as a catalyst to switch to paid, ad‑free plans, rival services, or open‑weight/local models, and argue that ad-funded AI agents pose deeper risks around manipulation, privacy, and even political influence.
Financial health and business model
- Many see the move as a bad sign for OpenAI’s finances, interpreting ads as “last resort” monetization and evidence of high burn and weaker-than-hyped revenues.
- Others argue it’s straightforward: huge free-tier costs require revenue; even highly profitable companies still chase “more,” so ads are expected, especially before an IPO.
- Some note that earlier public comments from leadership framed ads as a “last resort,” fueling accusations of inconsistency or dishonesty.
Where and how ads appear
- Ads are currently described as:
- Only for free-tier and very low-cost users, not Plus/Pro/Business plans.
- Clearly labeled and visually separated from answers.
- Contextual, based on chat content, history, and “advertiser-provided context hints.”
- Some users already report seeing ads earlier as part of A/B tests, sometimes irrelevant to their query.
User trust, enshittification, and model bias
- Dominant concern: this is the start of Google-style “enshittification” — ads will grow more intrusive and eventually blend into or bias answers.
- Even with explicit separation, many worry about:
- Stealth influence over rankings and recommendations.
- Training-time steering or slightly biased token weights in favor of advertisers or against competitors.
- Use of memories and sensitive history for ad targeting.
- A minority are fine with ads if they remain clearly marked and pay for a strong free service; a few even say they like well-targeted ads.
Competition and alternatives
- Several predict this strengthens the case for:
- Open-weight/local models without ads.
- Competing paid services that promise to remain ad‑free (e.g., other LLM vendors, alternative search engines).
- Others argue mainstream users won’t self-host, don’t care about open vs proprietary, and will accept ads like they do on Google.
Advertiser experience and economics
- Early advertisers report:
- Very high suggested CPCs (around a few dollars per click).
- Limited transparency, weak targeting controls, and poor initial ROI.
- Some speculate ad inventory and targeting quality are immature, but expect prices and sophistication to rise.