Our newsletterAcademicarXiv, 5 Aug 2026
Once ChatGPT started showing ads to an account, the median account saw an ad on 24% of its later prompts
Audience targeting
24%
of prompts showed an ad for the median exposed ChatGPT account after its first ad during March 8–31, 2026
- Reader Note:
- Based on 48 synthetic U.S. ChatGPT accounts that received at least one ad during March 8–31, 2026. Ad rate here is calculated per account as the fraction of prompts sent after that account’s first observed ad that surfaced an ad. This excludes accounts with no ads and excludes each account’s pre-exposure prompts.
- Evidence:
- Observed behaviour
- Vendor interest:
- None
- Sample:
- 91 synthetic ChatGPT accounts in the US, stratified in a 3×3 design by race/ethnicity signal (Black, Hispanic, White) and income tercile (low, medium, high), with at least 9 accounts per cell; 335 prompts; 127,801 conversations total.
Reader Note:
Evidence:
Vendor interest:
Sample:Good for
Comparing yourself against a similar segment, and choosing what to test next.
Not for
- Saying that one thing caused another, or predicting what you would get.
- This was measured in one market. It may not hold where you are.
- The two groups were not split at random — they picked their own side. The difference may have been there before anything was measured.
Related cards
Audience targetingTrend
TrendPrompts that name a brand are much more likely to show an ad
Audience targetingTrend
ChatGPT accounts in lower-income US ZIP codes were more likely to be shown ads
Idea
Go deeper on this topic
Our Substack turns the data behind these cards into full breakdowns: what changed, why, and what to do about it.
Audience targetingIdea
Category-adjacent ads became the norm at 58%