Editorial policy
Who writes here, what we test, what we don't, and who is accountable.
Who is accountable
Robin D. is the publisher and editor of bitcritiq. Every article published here — including everything drafted by AI — is reviewed by, and is the responsibility of, a named human being. If something on this site is wrong, that is a person’s fault, not a system’s.
Some of our writers are AI. Every byline says so.
We operate several AI analysts, each covering a subject area. They have names and profiles, and every byline they appear on carries the label “AI analyst” — not buried on a profile page, but next to the name, on the article, where you will actually see it.
We do this because the alternative has been tried. Several large publishers ran AI-written articles under invented human bylines, some with purchased headshots. In every case the damage came not from using AI but from the concealment. We would rather tell you up front.
Our AI analysts have no photographs, because they have no faces.
What AI writers may and may not do
This is the line that matters, and it is enforced in software, not by good intentions.
AI analysts do not write reviews. They do not write best-of lists, and they do not write head-to-heads. They never assign a score to anything. Those formats require having used the product, and an AI analyst has not used anything.
AI analysts do write research-led work: what a manufacturer’s own specification actually says, how a new product compares on paper against others in its price tier, what published measurements show, what owners and reviewers are reporting, and what a feature does. This work is legitimate precisely because it never claims experience it does not have.
Our publishing tool refuses to accept a scored article from an AI byline. It also refuses any AI-authored article containing phrases like “we tested” or “we measured”. These are not warnings — the article does not publish.
Showing our working
Every AI-authored article carries a “How we researched this” block stating why that subject was chosen, how it was researched, and — importantly — what the research cannot tell you. A spec sheet cannot tell you what a photograph looks like, and we say so rather than implying otherwise.
Every such article lists its sources, linked, with the date accessed. Where a manufacturer’s own documentation is the source, it is marked Official. Where a source is an academic, government or standards body, it is marked too. We would rather show you a short article with four real sources than a long one with none.
What human reviews mean
When an article carries a score, a person used the product. Every review states how long we had it, where the unit came from — bought, loaned, or supplied and kept — who paid for it, and what we could not test. That block is mandatory.
Corrections
We fix errors promptly and append a dated note. We do not silently amend a factual claim. Substantive corrections update the article’s modified date; the correction note stays permanently. Write to us and we will act within 48 hours.
Independence
We do not accept payment for a score or for a place in a best-of list. We do not give manufacturers preview of, or approval over, a verdict. We do not accept embargo terms that constrain what we can conclude.
Sponsored articles are permitted, are labelled Sponsored, are never scored, and never appear in our reviews section.
What we don’t cover
We do not publish financial, health, medical or legal advice. Those subjects demand credentials we do not have, and we would rather be silent than irresponsible.
Affiliate links
We earn commission on some links. Verdicts are written before any affiliate link is added, by people who cannot see what a link pays. We link the best price, not the best commission. Full disclosure.