Study plan, registered on 2 October 2026. This study has not been run yet. I’m publishing the method before any results exist so it can’t be adjusted to fit them. Results will be added to this page, with the date they were collected.
On 2 October 2026, 7 of the UK’s 10 biggest newsbrands blocked GPTBot and 9 blocked ClaudeBot, according to my October crawler study. This study will check whether the sites that block training crawlers appear in AI answers less often than the sites that don’t.
What will be measured
A fixed list of 30 UK news queries, covering politics, crime, sport and money, will be run on four AI engines. For each answer, each of the ten brands will be recorded as linked (a clickable link to its site), named without a link, or absent. The results will then be split by whether the brand blocked the matching training crawler on 2 October 2026.
The engines are ChatGPT, Perplexity, Claude and Google. Each query will be run logged out, in a clean browser profile, from the UK, on a single date. If an engine needs an account on the run date, it will be reported as not observed. It won’t be run under an account instead.
This measures whether brands appear in answers on one date. It won’t say anything about traffic.
The query list
The list below is fixed as of 2 October 2026. Every run will use these queries, worded exactly like this.
- who is the uk prime minister (politics)
- when is the next uk general election (politics)
- what was in the latest uk budget (politics)
- uk inflation rate latest (politics)
- nhs waiting list figures latest (politics)
- uk energy price cap october 2026 (politics)
- what is the two-child benefit cap (politics)
- uk immigration rules latest changes (politics)
- uk knife crime statistics (crime)
- how to report fraud in the uk (crime)
- uk shoplifting figures latest (crime)
- what is the sentence for dangerous driving uk (crime)
- county lines drug gangs explained (crime)
- uk prison overcrowding latest (crime)
- premier league table (sport)
- premier league results last night (sport)
- england football squad latest news (sport)
- f1 drivers championship standings 2026 (sport)
- wimbledon 2026 men’s final result (sport)
- the ashes latest score (sport)
- champions league results (sport)
- world cup 2026 final result (sport)
- bank of england base rate today (money)
- best savings accounts uk (money)
- state pension increase 2027 (money)
- uk house prices latest (money)
- is the uk in a recession (money)
- isa allowance 2026-27 (money)
- minimum wage increase april 2027 (money)
- energy bills support uk (money)
Limits
AI answers vary by date, location and wording, so the results will describe one day. Thirty queries on four engines can show a pattern, but the study won’t treat it as a general rule.
What would show the result was wrong
Running the same 30 queries, logged out, on the same day and getting clearly different counts for each brand.
Data
When the study runs, the results will be published here as a CSV, with one row per query and engine and a column for each brand.