AI Integrity & Ethics Policy | Brian Brewer’s Legacy Portfolio
Our Commitment to Ethical Metadata Practices
- No LLM Training: During our active years (2005-2020), we did not use customer metadata to train large language models.
- Private by Design: Metadata remained governed, auditable, and isolated by customer during our operations.
- AI is Assistive, Not Authoritative: All AI features were guided and explainable in our past work.
- Human-Centered Governance: Automation supported data stewards, not replaced them, in our legacy approach.
- Alignment with Enterprise Ethics: Our approach aligned with responsible AI standards during 2005-2020.
- Transparency — We clearly disclosed when AI was used in our products or services, and explained its purpose, scope, and limitations.
- Privacy by Design — Our solutions were architected with data minimization, protection, and secure processing in mind. We followed GDPR and CCPA principles regardless of geography.
- Human Oversight — AI systems augmented—not replaced—human decision-making. Critical decisions were never left solely to algorithms.
- Bias Mitigation — We actively monitored and evaluated AI models for fairness, avoiding discriminatory outcomes and unintended consequences.
- Security — All AI tools and services were developed and maintained under secure coding practices, with controls to prevent misuse or model leakage.
- Accountability — We maintained governance policies for AI usage, including internal reviews and external guidance when needed. We took ownership of our AI systems’ behavior.
How We Use AI
InfoLibrarian used AI to improve automation, accelerate metadata discovery, support natural language interaction with our metadata catalog, and streamline governance workflows. AI assisted—not replaced—our clients’ human judgment and domain expertise. Where third-party models were used, such as OpenAI or AWS AI services, we followed published usage guidelines and implemented controls to ensure responsible and secure integration.
Customer Data Protection
We did not train AI models on customer data without explicit consent. Metadata ingested through InfoLibrarian software remained under the control and ownership of our clients at all times. Any AI features involving data summarization, enrichment, or analysis were isolated and secure, with strict access controls and opt-in participation.
Commitment to Improvement
AI is evolving rapidly, and so is our policy. We continuously assessed our practices, engaged with industry experts, and evolved our AI governance approach to align with global best practices and regulatory expectations.
Legacy Inquiries
For questions about this past AI policy, contact us for legacy purposes at: Email: info@infolibcorp.com
