Humanit uses AI to assist with writing, analysis, and originality checking. AI does not replace human judgement - outputs may contain errors and should be reviewed before use in important decisions. Originality and AI-likelihood scores are estimates, not proof. Humanit is designed to improve writing, not to facilitate deception. We are committed to transparency, fairness, privacy, and continuous improvement.
Humanit uses artificial intelligence and machine learning models to provide writing assistance, text analysis, and originality checking.
Specifically, AI is used to: analyse text for AI-likelihood and human-written characteristics; detect potential plagiarism patterns; assess academic readiness; correct grammar, spelling, and punctuation; rewrite and paraphrase text in various styles; generate Writing Signature profiles for personalised output; and produce quality and change summaries.
AI is a tool that supports these functions. It processes text you submit during your active session and returns results for your review and use.
The AI models used by Humanit are provided through third-party service providers and are selected based on their performance, reliability, and data protection practices.
AI is a powerful assistant, but it is not a substitute for human judgement, expertise, or critical thinking.
The outputs, suggestions, and analyses produced by Humanit's AI tools should be treated as drafts or reference points - not as final, authoritative, or professionally certified content.
Human judgement is essential for:
• Determining whether content is accurate, appropriate, and suitable for your specific context.
• Ensuring the tone, meaning, and intent of text align with your goals.
• Making decisions based on originality scores, AI-likelihood indicators, or plagiarism assessments.
• Verifying that outputs comply with applicable laws, institutional policies, and ethical standards.
You remain responsible for all decisions made based on or influenced by Humanit's AI outputs.
AI-generated and AI-analysed content is inherently probabilistic. Outputs may contain factual errors, omissions, stylistic inconsistencies, or contextually inappropriate suggestions.
This is particularly important when outputs are used for significant decisions, including but not limited to:
• Legal decisions: AI outputs are not legal advice and must be reviewed by a qualified legal professional.
• Financial decisions: AI outputs are not financial advice and must be verified before use in financial contexts.
• Medical decisions: AI outputs are not medical advice and must never replace professional medical guidance.
• Academic decisions: AI outputs must be reviewed for accuracy, originality, and compliance with academic integrity policies.
• Employment decisions: AI outputs must not be the sole basis for hiring, evaluation, or disciplinary decisions.
• Regulatory decisions: AI outputs must be reviewed for compliance with applicable regulations and standards.
You should always review, edit, and verify AI outputs before relying on them for any important decision.
Humanit provides originality scores, AI-likelihood percentages, plagiarism indicators, and academic readiness assessments. These are analytical estimates based on pattern recognition and statistical modelling.
These indicators are not definitive proof of any characteristic of the text. They have inherent limitations and may produce false positives or false negatives.
Different detection tools use different methodologies and may produce different results on the same text. Humanit's indicators should not be treated as conclusive evidence.
These estimates are provided for informational purposes only and should not be used as the sole basis for any judgement about authorship, originality, or academic integrity.
Humanit is designed to help users improve their writing, verify originality, and produce clearer, more effective text.
The platform is not intended to be used as a tool for deception, fraud, or academic dishonesty. This includes, but is not limited to:
• Using the service to deliberately circumvent academic integrity policies or institutional rules.
• Using the service to misrepresent AI-generated content as entirely human-written for fraudulent purposes.
• Using the service to plagiarise or pass off another person's work as your own.
• Using the service to generate misleading or deceptive content for financial, legal, or professional gain.
Users are responsible for ensuring their use of Humanit complies with applicable laws, institutional policies, and ethical standards.
We encourage users to use Humanit transparently and in good faith, as a writing improvement tool rather than a tool for misrepresentation.
Transparency: We are open about how we use AI, what data is processed, and the limitations of our tools. This Responsible AI Statement is part of that commitment.
Fairness: We strive to ensure our AI tools work equitably across different writing styles, languages, and contexts. We continuously monitor for bias and work to reduce it.
Privacy: Humanit minimises retention, uses private processing access for signed-in uploads, and keeps account content only where a feature explicitly saves it or the user chooses to save it. See our Privacy Policy for full details.
Continuous Improvement: We actively seek user feedback, monitor AI performance, and refine our models and processes to improve quality, accuracy, and reliability over time.
We are committed to responsible AI development and use, aligned with emerging best practices and regulatory expectations.
Effective Date: 1 July 2026
Last Updated: 5 July 2026
Version: 1.0
Version History:
• v1.0 (1 July 2026): Initial publication of the Responsible AI Statement.
We will update this statement as our AI practices evolve. Material changes will be reflected in the version history and the Last Updated date.
How can we help?