TAC219 General Information

Purpose and Scope

TAC 219 establishes a mandatory AI Code of Ethics for Texas state agencies and local governments. It promotes responsible innovation while ensuring that AI systems remain trustworthy, lawful, and aligned with public values. It also aligns AI governance with the NIST AI Risk Management Framework.

Ethical Principles of Artificial Intelligence

Human Oversight and Control

AI systems must operate under appropriate human oversight. Humans must be able to review AI inputs and outputs and disable systems that produce harmful or inaccurate results. Higher-risk systems require enhanced oversight.

Fairness

AI systems must not infringe on protected rights or cause unlawful harm. Agencies must ensure representative data and strong data governance to maintain equitable outcomes.

Accuracy

Agencies must ensure AI system outputs are accurate. Employees should validate results, and agencies must monitor accuracy throughout the system lifecycle.

Redress

Agencies must provide mechanisms for individuals to challenge harmful AI decisions. A designated contact and internal review procedures are required.

Transparency

Agencies must understand and disclose AI use in public-facing and consequential contexts. AI systems must not be presented as human.

Data Privacy

Agencies must protect personal identifying information (PII) through data minimization and employee training to prevent misuse in public AI tools.

Security

AI systems must be protected from threats including data poisoning, unauthorized access, and model exfiltration. Vendors must disclose vulnerabilities promptly.

Accountability and Liability

Agencies remain responsible for AI outcomes. Staff must be trained, vendors held accountable, and legal and records requirements followed.

Evaluation and Documentation

AI systems require ongoing evaluation to ensure continued public benefit. Agencies must maintain documentation of system performance, data sources, and lifecycle modifications.

Rule 219.11: Code of Ethics and the Ethical Principles of Artificial Intelligence