Spelt with an s, because that is how the search is typed here. An AI humaniser UK search usually comes from one of two places: a student at a UK institution whose submission has been flagged, or a freelance writer whose client has run a detector over an invoice-ready draft.
The UK case has one feature worth naming up front. Academic misconduct here is a formal process with published procedures and a right of reply, which changes what is worth doing when you are accused. A rewrite is not it.
The procedural shape matters more than the detector score, and it is the part tool pages never mention.
UK institutions generally handle suspected unfair academic practice through a documented procedure: an allegation, disclosure of the evidence, a chance to respond, and a panel or interview. That is a slower and more evidence-driven process than a single score, and it's the reason a rewrite is the wrong reflex.
What carries weight in that setting is a record of how the work was produced. Version history, dated drafts, your notes and reading, and the ability to talk through the argument in person. A rewritten draft is the one thing that destroys all of it at once.
It also helps to understand what the detector figure is worth. Turnitin publishes about 4% false positives across the full distribution and under 1% at document level above its 20% reporting threshold; Weber-Wulff et al. (2023) measured 4% to 12% independently. Across a 200-student cohort, 4% is roughly eight human-written submissions flagged. Our Turnitin explainer covers those figures and the threshold guidance in full.
If you write English as an additional language, bring the research too. An average false positive rate of 61.3% across seven detectors on TOEFL essays by non-native writers, their measurement rather than ours between them (Liang et al., 2023); native-written controls in the same study were barely flagged.
The procedure gives you a right of reply, which changes what is worth doing with the time.
Because UK unfair-practice processes disclose the allegation and invite a response, you usually have a window in which evidence is more useful than anything else you could produce. A rewritten submission is the one thing that closes that window, since it replaces the text under examination and overwrites the revision history behind it.
Request the evidence. Read your institution's procedure. Assemble the record: revision history, dated drafts, reading notes, the outline, and your ability to talk through the argument at interview. What carries weight is set out here.
Before submission the calculus is different and a rewrite is ordinary editing. The free path gives 3 humanizations a day at up to 260 words with no account, or 3 actions a day (scans or humanizations, any mix) at up to 260 words per humanization on a free account.
Work one paragraph at a time. It is the only scale at which you'll genuinely read each change, and on a dissertation chapter reading each change is the entire point.
Use Light wherever the wording is load-bearing: it varies sentence rhythm, cuts connective filler and leaves your vocabulary where you put it. Balanced edits phrasing more broadly. Maximum replaces most of the wording and is a paid mode. Then check the sentences carrying references, because a softened claim detached from its source is a citation error you introduced. The check is here.
We can't tell you whether a rewrite clears a particular UK institution's checker. No vendor sees another system's verdict, and those systems retrain against rewriting techniques, so any guarantee is marketing rather than a measurement.
We score before and after with our own detector, labelled as ours, and highlight the words that changed. That is a real signal and not a prediction.
The detector is built and measured for English, so it reads British English as English. It doesn't score other languages and we won't imply it does.
Both refer to the same kind of tool. British English favours the -iser spelling and that is how the search is usually typed in the UK, so this page uses it throughout; American English uses -izer. There is no product difference behind the two spellings, here or anywhere else. What does differ between tools is whether you can see what the rewrite changed.
Nobody can promise that and you should distrust any vendor who does. We cannot see what your institution’s checker outputs, and those systems retrain against rewriting techniques. More importantly, UK academic misconduct is a documented procedure with a right of reply rather than an automatic verdict, so the useful preparation is evidence of how you wrote the work, not a smoother draft.
Do not rewrite the submission. Request the evidence, read your institution’s unfair academic practice procedure, and assemble your process record: version history, dated drafts, notes, reading list and your outline. Then cite the published false-positive figures, because roughly 4% across a large cohort means human-written work does get flagged and the procedure exists precisely for that.
Turnitin’s own 2024 documentation gives about 4% across the full distribution and under 1% at document level above its 20% reporting threshold, and it tells instructors not to act below 20%. Weber-Wulff et al. (2023) measured 4% to 12%. If you write English as an additional language, Liang et al. (2023) measured an average false positive rate of 61.3% across seven detectors on TOEFL essays by non-native writers.
The rewriter follows the register of what you give it, so a draft written in British English generally comes back in British English. It is not a spelling converter and we would not sell it as one. If spelling consistency matters for a submission, check the output, because a rewrite is a drafting aid rather than a style guide.
Ours gives 3 humanizations a day at up to 260 words each with no account, and 3 actions a day (scans or humanizations, any mix) at up to 260 words per humanization on a free account in Light and Balanced modes. Maximum is a paid mode. Scoring your own text costs nothing extra and is usually the more useful half, particularly if you are trying to understand a flag rather than change a draft.
Yes, and in the UK procedural context that is the main risk. Pasting a rewrite over your working document makes the revision timeline show a large block of text arriving at once, which is the opposite of what you want it to show a panel. Rewrite a copy and keep the original history intact.
Score a draft in British English before anyone else does.
See the detector →The detector-side ranking for UK users.
See the ranking →The process record that carries weight in a formal procedure.
See the evidence list →The figures, the 20% threshold, and what they mean.
Read the explainer →The citation check before you submit anything.
Read the method →Policy first, then the tool.
Read the guide →Score a paragraph and read exactly what changed, on a copy rather than your submitted file. 3 scans a day at up to 1,000 words, no account, no card.