Collin College's proposed AI policy would require human review of AI-made content and bar student records from chatbots

Collin College employees could use artificial intelligence on the job, but only approved tools, with a person checking the results and without feeding in student or personnel records, under a new policy up for a vote Friday, Oct. 30.

The rules, filed as policy CRB (Local), are on the board of trustees' agenda for a second reading and approval at its regular meeting at the Collin Higher Education Center in McKinney.

A memo from General Counsel Monica Velazquez says the board's Organization, Education and Policy Committee is recommending the policy unchanged, and that it grew out of the college's review of all its policies with recommendations from the Texas Association of School Boards' policy service.

Who decides which tools are allowed

The policy applies to AI used on college devices and on personal devices used for college business. The district president or a designee would set up a standing AI Adoption & Innovation Committee to approve AI systems, assess risk and publish a list of tools cleared for general use. Only tools on that list should be used for college business.

The college's chief innovation officer would be responsible for compliance procedures, and the college could monitor AI use on college-issued devices.

The draft names ChatGPT, Gemini, NotebookLM and Copilot as examples of generative AI. It sets stricter review for a "heightened scrutiny AI system," one meant to make or heavily drive decisions such as admissions, financial aid, hiring, discipline or access to services the college must provide.

Rules for AI output

AI-generated content would have to be reviewed by qualified staff before it is published, shared with decision makers or used in college business, with the review documented. Employees would be responsible for inaccurate or fabricated AI content, which the policy calls "hallucinations," that they publish without that review. Systems would have to be tested before deployment, though internal drafts are exempt.

Users are told to assume anything typed into an AI tool may be stored, used to train models or made public. The policy bars entering Social Security numbers, financial account data, student records protected by federal privacy law, confidential personnel files, health records, passwords and other confidential information without written permission.

Prohibited uses include violating someone's constitutional rights or causing unlawful harm, creating unlawful deepfakes, helping commit crimes such as phishing, and getting around the college's security controls. When AI is used for college business, the committee is to set disclosure standards requiring clear, plain-language notice in writing.

Training and risk rules

The college would adopt a nationally recognized AI risk framework, such as the National Institute of Standards and Technology's AI Risk Management Framework, and offer training on permitted uses, data safeguards and reporting violations. Citing state law, the policy says employees who use a computer for at least a quarter of their required duties must complete annual AI training.

Employees and students who use AI tools to deceptively harm, bully or harass others or to make unlawful deepfakes would face discipline.

The same agenda item updates the employee grievance policy to cover complaints about faculty rank and revises the process for complaints from the public.

Sources

meetings.boardbook.org

meetings.boardbook.org

meetings.boardbook.org