AI Bias Report — Tarosyn
This page presents Tarosyn's periodic report on AI fairness, our bias audit process, findings, and remediation actions.
Audit Approach
We review AI reading outputs across a sample of zodiac signs, life situations, and question types to check for systematic variation in tone, length, or positivity that could indicate demographic or cultural bias. Audits are conducted quarterly.
Known Risks
- Training data for large language models may over-represent Western esoteric traditions. We actively prompt for balanced treatment of Eastern and indigenous traditions.
- AI tone may vary based on question framing; we monitor for harmful differences in responses to emotionally sensitive queries.
Remediation
Where bias is identified, we update the system prompt instructions used for reading generation. Results are validated in the next audit cycle.
Contact
If you notice AI outputs that appear biased or culturally insensitive, report them to [email protected].