humanise.ai editorial mark

Who to email
about what.

No support desk. These addresses go to people who read them.

We don't run a support desk. These addresses go to people who read them.

General

hello@humaniseai.ai - questions, feedback, partnership enquiries.

Security

security@humaniseai.ai - vulnerability disclosure. See Security for scope and safe-harbour terms, or the machine-readable version at /.well-known/security.txt.

Privacy and data requests

privacy@humaniseai.ai - GDPR / UK-GDPR / CCPA requests. We respond within 30 days.

Abuse

abuse@humaniseai.ai - reports of misuse of the humaniser.

Policy

policy@humaniseai.ai - institutional questions about AI writing policy, detector evaluation, or academic integrity.

Press

press@humaniseai.ai - we respond to press enquiries within 2 working days.

What to put in a message so we can act on it quickly

Most of the delay in reading an email is working out what the sender actually needs. If you want fast turnaround from us, write the first line as a one-sentence summary of the ask: "This is a vulnerability report," "I'm a journalist with a deadline," "I run a university and want to discuss policy." Then put the detail underneath.

For security reports, include a minimal reproduction, the HTTP response you observed, and the time (with timezone) of the request. We keep limited, anonymised operational logs for a short period, so a timestamp inside that window helps us correlate your test to the live record even though we never store the text you sent.

For privacy requests, state the right you are exercising (access, deletion, portability, rectification, or objection), the jurisdiction you want it processed under, and enough context for us to tie your request to any record we might hold. Because the humaniser is stateless and we do not store submission text, in practice most of what we hold under your name is log metadata and any correspondence you have sent us.

For policy questions from schools, universities and employers, include the specific decision you are being asked to make. We will give you a short written note you can forward internally rather than a marketing pitch, and we will tell you where the evidence is weakest.

What not to send

Please do not send us copies of other people's writing to "check" for AI content. We do not run the detector tools we review here, and we do not consult on specific student or employee cases. If you are in that situation, read the AI detection explained guide first and treat any single detector score as a flag worth investigating, not as proof.

If something urgent is on fire and the humaniser itself is down for all users, check the status page before emailing. If the status page is also down and this is during UK working hours, hello@ is the fastest contact. Outside UK working hours, security@ is monitored for urgent live-site issues only; please do not use it as a general inbox.

Response-time expectations by inbox

These are the targets we aim for and the realistic variance around them. If an email misses the target consistently, that is a signal our side needs more hands, not a signal to shout louder on the next thread.

  • hello@: 1 to 2 UK working days for a substantive reply. If you have asked a specific technical question that someone can answer in a sentence, often same-day. We do not use templated replies.
  • security@: within 2 UK working days to acknowledge, with a follow-up time estimate for triage. For a confirmed exploitable vulnerability, we aim to ship a fix within one hour of acknowledgement, as noted on the security page.
  • privacy@: up to 30 calendar days for a formal response, as the UK GDPR and CCPA require. Simple requests (access, erasure confirmation) are usually answered within 5 working days.
  • abuse@: within 2 UK working days to acknowledge an abuse report, with the specific action we have taken or will take. If you report misuse of the humaniser against a named victim, we treat it as a priority.
  • policy@: within 3 to 5 UK working days for a written note that is forwardable inside your organisation. We do not sell sales calls.
  • press@: within 2 UK working days; we will respond at interview length if you need quotes for a piece on a deadline.

If you want to send us a correction

We make mistakes. The blog is written against our best current understanding of the detector market, and the market moves. If a figure on this site is out of date, a vendor has changed its pricing, a cited paper has been withdrawn, or we have described a detector's behaviour in a way that no longer matches reality, send the correction to hello@ with the URL of the page and a one-line description of what is wrong. We will acknowledge, verify, and update in place with a dated note at the bottom of the piece. We do not quietly rewrite the past, and we do not sit on corrections to protect our ego about the first draft.

Corrections about AI policy, security and privacy go through the specific inbox for that area, because the correction is itself a formal statement and we want it logged in the thread for the right audit trail.

What this contact page is deliberately not

It is not a lead-capture form that routes to a sales team because there is no sales team. It is not a chatbot pretending to be a person. It is not a ticket system with a wait estimate that is wrong on principle. Those three design choices are deliberate. The pricing page explains the economic logic that makes a stripped-down contact surface possible; the about page explains why we think the stripped-down version is a feature and not a cost cut.