Template — must be reviewed by qualified legal counsel for your jurisdictions (e.g. Canada/PIPEDA, provincial laws, GDPR, US state AI-in-hiring rules) before launch. Not legal advice.

AI processing disclosure

Last updated: [DATE]

1. What is sent to the AI provider

When you confirm an AI operation, SkillMatch AI sends the relevant content to Anthropic's Claude API: the extracted CV text (or the profile details you entered), the job description and any requirements you curated, and our analysis instructions. Nothing is sent before you confirm the estimate. Candidate names may appear inside CV text; the application refers to candidates by reference where possible.

2. How the provider handles data

Anthropic processes API data under its commercial terms and data-retention policy. The operator of this service is responsible for reviewing and configuring those terms (including any zero-data-retention or regional options). SkillMatch AI does not use your documents to train models.

3. What the AI is instructed to do — and not do

  • Ground every statement in the supplied documents, quote evidence, and mark unknown or ambiguous information as such.
  • Never treat missing information as proof that a candidate lacks a skill.
  • Never infer protected or sensitive characteristics (such as age, gender, ethnicity, religion, disability, health or family status), and never score personality, appearance or “culture fit”.
  • Never recommend hiring, rejecting or ranking candidates.

4. Limitations

  • AI output can be incomplete or wrong, and can misread poorly formatted or scanned documents.
  • Reports reflect only the documents supplied and are not a measure of ability or future job performance.
  • Token estimates are shown before processing; final usage may differ within the reserved maximum.

5. Human review

A person must review every report and make every hiring decision. Recruiters can correct extracted data, challenge findings, and verify uncertain points using the interview guide.

6. Your obligations as a customer

  • Have a lawful basis to process each candidate's data with an external AI provider, and give candidates any notice (and obtain any consent) required in your jurisdictions.
  • Comply with laws on automated decision-making and AI in employment that apply to you (for example notice, bias-audit or opt-out requirements).
  • Respond to candidate requests to access, correct or delete their data; the application provides export and deletion tools.