Data Use Policy

Sourcing

Our datasets are built from licensed sources and synthetic generation. We do not scrape personal data or use data without the rights to do so.

De-identification

Records are de-identified following the HIPAA Safe Harbor method where applicable: names, addresses below state level, dates more specific than year where required, contact details, and account numbers are removed or generalized. Synthetic records contain no real individuals.

Permitted use

• Training, fine-tuning, and evaluating AI models under your license.
• Internal research and model risk assessment.

Prohibited use

• Re-identifying or attempting to re-identify any person or property owner.
• Combining data with other sources to profile individuals.
• Using data to make unfairly discriminatory decisions, including on protected characteristics.
• Redistributing data outside your license.

Provenance and integrity

Every release carries version, lineage, and integrity metadata so you can document training data for audits, including EU AI Act Article 10 data governance requirements.

Reporting concerns

If you believe a record contains personal information or violates this policy, contact cyberprosoftware@gmail.com and we will investigate and remove it where appropriate.