The Texas Higher Education Coordinating Board (THECB) has adopted a code of ethics governing how the agency procures, develops, deploys and uses artificial intelligence.
The policy, approved during a July 22 board meeting, is intended to ensure that AI is used effectively and ethically within the coordinating board. It was adopted under a state law requiring agencies to implement the AI code of ethics created by the Texas Department of Information Resources.
THECB adopted an abbreviated, agency-specific version of DIR’s code, preserving its core principles while adding agency complaint procedures and omitting many of DIR’s detailed implementation requirements.
The framework addresses human oversight, fairness, accuracy, redress, transparency, data privacy, security, accountability, evaluation and documentation. It also makes clear that governmental entities remain responsible for decisions and outcomes when tasks or decision-making functions are delegated to an AI system.
Human control is intended to reduce the risk of inaccurate or undesirable outputs and allow agencies to revise a system’s rules and supporting data. AI systems also should not prevent people from reviewing their decisions.
The agency is also directed to consider whether the data used to develop a system adequately represents the people or subjects affected by its judgments, decisions or predictions. AI-generated conclusions should be evaluated within the context in which they are used because inaccurate outputs remain an ongoing risk.
Members of the public may submit complaints about the agency’s use of AI to its privacy office or through the existing student complaint portal. Providing a method for redress is intended to support public trust in the agency’s use of the technology.
Transparency may include providing information about a system’s data, models and outputs. Preliminary assessments, monitoring, testing and complaints are among the records that can help the agency understand a system throughout its life cycle.
Privacy provisions emphasize limiting the personally identifiable information entered into AI systems. Using personal information in tools outside a secure government environment could violate privacy laws or the agency’s obligations to the people whose information is involved.
Security risks identified in the framework include malicious code injection, data poisoning, unauthorized access and the removal of data or models from an AI system. Those risks are to be considered throughout the system’s life cycle.
The document defines a heightened scrutiny AI system as one that autonomously makes or serves as a controlling factor in a consequential decision affecting a person’s access to a government service. It also defines an AI risk officer as the person responsible for overseeing the agency’s management of risks associated with those systems.
The requirements apply to the coordinating board as a state agency. The agenda materials do not impose new requirements on Texas colleges and universities.
Texas Higher Education Board Adopts AI Code of Ethics
What to Know:
- The policy creates a formal process for complaints and review when the agency’s use of AI produces an adverse outcome.
- It emphasizes human review, privacy, security, transparency and agency accountability.
- The policy governs the coordinating board, not Texas colleges and universities.
