Acceptable Use Policy (AUP)

The short version

• Don't use deepaidetector to hurt people.
• Don't use a single AI-detection score to fire, fail, expel, or accuse anyone — combine with human review, give the subject a chance to respond, follow your local due-process law.
• Don't abuse our infrastructure (scraping, scripting around rate limits, reselling).
• Don't submit illegal content.
• Don't try to break our detection model or bypass our paywall.

1. Why this policy exists

AI text detection is a sensitive technology. False positives — real human writing flagged as AI — are a real risk and have real consequences when used in academic, employment, or moderation decisions. This policy sets the boundaries we expect every customer to operate within, both to comply with law and to keep the Service trustworthy.

This AUP is part of the Terms of Service. Breaching it can result in suspension, termination, refund forfeiture, and (in serious cases) legal action.

2. Decision-making rules (AI-detection-specific)

If you use a detection score as input to a decision about a person, you must:

• Not rely on the score as the sole basis for any disciplinary, academic, employment, hiring, firing, immigration, credit, insurance, housing, criminal-justice, or content-moderation decision.
• Combine the score with human review by a qualified person who looks at the underlying text and context.
• Disclose to the subject that an AI detection tool contributed to the analysis, where required by law (EU AI Act Art. 50, certain US states, school policies).
• Provide a right to respond — let the subject see the verdict, contest it, and supply evidence (drafts, edit history, voice samples) before any adverse outcome.
• Honour your local due-process obligations (eg, university honour-code procedures, EU AI Act high-risk-system obligations if applicable, employment law).
• Not assume the score equals certainty. See our AI Detection Disclaimer.

3. Prohibited content

You may not submit, store, transmit, or otherwise process via the Service any content that:

• Depicts or sexualises minors (CSAM). We report CSAM to NCMEC where required.
• Promotes terrorism, mass violence, or genocide.
• Incites imminent violence against a person or group.
• Constitutes non-consensual intimate imagery or doxing.
• Infringes third-party intellectual property without authorisation.
• Discloses third-party confidential information you have no right to disclose (trade secrets, classified data, attorney–client material, sealed court records, patient records).
• Contains malware, exploits, or material designed to compromise systems.
• Is otherwise illegal under any law that applies to you or to us.

4. Prohibited uses

You may not use the Service to:

• Harass, threaten, intimidate, stalk, defame, or discriminate against any person or group on the basis of any protected characteristic.
• Run any operation whose purpose is to harm specific individuals by mass-classifying their writing.
• Generate evidence to support a knowingly false accusation.
• Conduct legally regulated profiling (eg, EU AI Act prohibited practices) — see AI Act disclosure.
• Operate a service for governments to surveil or suppress political speech, dissent, journalism, or whistleblowing.
• Operate a service to pre-screen job applicants without meeting all applicable employment-discrimination, EEOC, EU AI Act high-risk, and similar obligations.
• Re-identify pseudonymous authors of writing (no de-anonymisation operations).

5. Technical / infrastructure rules

You may not:

• Bypass rate limits, paywalls, or quota systems (eg, rotating accounts, rotating IPs, splitting documents to dodge size limits).
• Scrape, crawl, or index the Service without our written permission.
• Resell, white-label, or syndicate the Service or its outputs except under an Enterprise reseller agreement.
• Reverse engineer, decompile, disassemble, or attempt to derive the detection model's weights, training data, or rule set.
• Probe, scan, penetration test, or load-test the Service except under a coordinated vulnerability-disclosure agreement (email security@deepaidetector.com first).
• Use the Service to launder traffic, proxy other services, or relay messages on behalf of third parties.
• Interfere with our infrastructure, sub-processors, or other customers.
• Share API keys, embed them in client-side bundles, or otherwise expose them.
• Operate a competing AI detector and use our outputs as training labels.

6. Adversarial-research carve-out

We support security research and academic study of detector robustness. If your research involves probing, fuzzing, or adversarial testing, contact security@deepaidetector.com in advance. Coordinated disclosure under our security.txt is welcomed and rewarded.

7. Free-tier specific rules

• One account per natural person on free tiers. Sock-puppet accounts are forfeitable without refund (none is due — it's free).
• Anonymous tier is rate-limited per IP hash. Rotating IPs to defeat the limit is a violation.
• Programmatic / scripted use of the anonymous or registered-free tier is prohibited. Use the API on a paid plan.

8. Enforcement

We may, at our discretion:

• Warn you and ask you to stop.
• Suspend or terminate the account.
• Forfeit any unused PAYG balance for severe violations.
• Cooperate with law-enforcement requests that meet legal standards.
• Pursue civil or criminal remedies.

For non-emergency violations we will use reasonable efforts to give notice and an opportunity to cure. Emergency suspensions (CSAM, active attack, court order) are immediate.

9. Reporting violations

To report abuse of the Service by another user, email abuse@deepaidetector.com. Include URLs, account names if known, screenshots, and a description. We respond within 5 business days; CSAM and credible imminent-harm reports get same-day response.

10. Changes

We may update this AUP. Material changes will be announced via email and the dashboard with at least 30 days' notice.