AIDramaHome›Editorial policy
Editorial policy
Who writes these reviews, how a score is reached, what we do when we are wrong, and what we refuse to do. This page exists so you can judge the judgement.
Who writes this
Oleksandr Korop is the named publisher and is responsible for everything on this site, at the address on the imprint. Reviews are written and verified by him with the AI Drama Reviews editorial team. There is no anonymous byline anywhere on this site and no content is published under an invented author name.
How a review is produced
- The app is installed on a United States account, on both iOS and Android where both exist.
- It is watched until the free tier runs out, so the free route is measured rather than assumed.
- At least one paywall is paid through with our own money, and the receipt is kept.
- Store listings, prices and dates are read directly and recorded with the date observed.
- Scores are assigned on the six published axes, and the weighted mean is published beside the verdict.
- Every figure in the finished page carries a source and a date.
What counts as evidence, and what does not
| Kind of claim | How we treat it |
|---|---|
| A price on a store listing | Quoted with the date observed. Treated as the strongest available evidence. |
| A company statement about its own production | Quoted verbatim and attributed. We do not paraphrase an AI claim. |
| A third-party crawl or estimate | Published with the source named and the method flagged as modelled, not audited. |
| A pattern in user reviews | Reported as a user claim, never as a finding, and labelled as unverified. |
| Something nobody publishes | Written as "not published". We do not estimate into a gap and present it as a number. |
The rule we apply to AI claims
No public information is not the same as no AI, and we never write it down as one. Four of the sixteen apps we rank have published something about AI in production. The other twelve have published nothing. On those twelve we record an absence of information, not an absence of AI, and our own catalogue judgement is labelled as judgement, dated, and revisable.
Do you use AI to write these reviews?
AI tools are used for drafting, structuring and checking consistency across pages. Every factual claim is verified by a human against a primary source before publication, every score is assigned by a human, and the named publisher is accountable for the result. No page is published that a person has not read end to end.
Corrections
We correct in public. Send a correction with a source and, if it holds, we change the page, change the "last verified" date and say what changed. We do not quietly edit. Where a company disputes a figure but publishes nothing to replace it, we say that too.
What we will not do
- Sell a position in the ranking, a score, a badge or a link.
- Publish a review of an app nobody has installed.
- Present an estimate as a measurement.
- Remove a criticism because a company asked.
- Describe an app as AI-free because it has not said otherwise.
How often this is re-checked
Prices and store data are re-verified at least every 90 days, and sooner when a reader reports a change. Each review shows the date it was last verified rather than a vague "updated" stamp.