Last updated: August 10, 2026
Responsible AI Use Policy
AI can assist work. It does not replace responsibility for facts, rights, judgment, or consequences.
Human responsibility
A person remains responsible for the purpose, inputs, review, decision, communication, publication, and deployment of AI-assisted work.
Check the work
Verify facts, sources, calculations, citations, dates, names, rights, and assumptions. Treat confident language as unverified until checked.
Protect information
Do not enter confidential, personal, regulated, proprietary, or security-sensitive information into an AI service without authorization and an approved use.
Bias and unequal effects
Consider who may be excluded, misclassified, stereotyped, disadvantaged, or exposed. Test with relevant examples and involve affected people where appropriate.
High-impact decisions
Do not rely solely on AI for decisions with legal or similarly significant effects. Use qualified human review, documented criteria, appeal or correction pathways, and applicable legal safeguards.
Document the workflow
Record the goal, context, data source, instructions, model or tool where relevant, review points, tests, known limits, and final human decision.
Third-party tools
The Lab may use OpenAI products and Codex. SozoRock AI Lab is not affiliated with, sponsored by, or endorsed by OpenAI. Participants must follow the provider’s current terms, privacy settings, and organizational rules.
Report problems
Report harmful, discriminatory, unsafe, confidential, deceptive, or security-related output through the Security and AI Incident page. Preserve relevant evidence without sharing additional sensitive information.