A new draft policy proposal discourages Wake teachers’ use of artificial intelligence detectors, requires students to acknowledge and explain their use of AI, and elaborates in greater detail on how AI may be used ethically.
The four-page policy elaborates on skills the district wants students and staff to learn and standards the district wants them to meet. While principals can make decisions about which AI tools will be used in their schools, Superintendent Robert Taylor said there are expectations for what students and staff learn about AI.
“We’ve clearly said this is an expectation; they have to do it,” Taylor said. “They can’t just choose not to do that.”
Board Member Jennifer Job said she’d heard from some students that they are afraid pushback against AI will take it out of their schools and leave them behind for life after high school.
“They want to learn this stuff,” Job said.
The draft policy presented Wednesday is much more detailed than earlier versions of a policy and than adopted policies around the Triangle and surrounding region.
Development of an AI policy is one part of addressing a technology already disrupting education in Wake’s classrooms and classrooms across the world, leading to fears of stifled learning and disengagement.
The school board’s policy committee decided without a formal vote Wednesday to move forward with the draft policy, while some board members suggested changes. Those include more details on what types of AI use could be violations of the code of conduct — such as altering images in an attempt to bully — and more details on oversight of AI use.
Board Chairman Tyler Swanson said AI is evolving and many people might need clarification that AI misuse isn’t just academic.
District officials said they believed many ways in which AI could be used to break conduct policies would already be covered without spelling it out too much.
“Cheating is cheating, plagiarism is plagiarism,” said Julie Crain, the district’s senior director of policy.
Job said she also wants it to be clear that teachers can’t use AI as the final or only means of grading an assignment. She knows some teachers have developed tools to assess student work.
Committee Chairwoman Lynn Edmonds encouraged board members to take the policy to their board advisory committees for further feedback, before the full board discusses a final version in August or September.
Two parents spoke in favor of a strong AI policy during the board’s public comment later Wednesday night.
Parents should have more information, and the policy should clearly spell out what data will be input into AI tools and how long AI tools will retain that data, said Becky Lew-Hobbs, the chair of the Wake County chapter of Moms for Liberty. She added that the policy is a good foundation that understands the risks posed by AI.
Parent Isabel Heineman said the policy needs to more clearly address how AI use will add to students’ learning, rather than replace it. Students need to learn, master skills, hone their creativity and make sound judgments, she said.
The board won’t vote on the policy until likely August or later, which, by rule, would require approval at two separate meetings.
The board and school district haven’t moved quickly to adopt an artificial intelligence policy, largely sticking with internal guidance while most other school districts adopt the same stock few paragraphs committing to AI literacy and academic integrity.
But whatever the school board may end up adopting could need to be revised in the near future.
House Bill 301 would require the State Board of Education to adopt age-appropriate standards for instructing students on AI literacy, developed by the Friday Institute at North Carolina State University. It would then approve a model AI policy for local school boards like Wake to adopt, though they wouldn’t be required to adopt it word-for-word.
That model policy would touch on topics now covered in Wake’s latest draft, including measures for data privacy and security and standards for AI use.
Different versions of the bill have passed the House and the Senate, and the Senate’s version — the latest version — is now before a House committee.
Publicly discouraging the use of AI detectors
Wake’s new draft AI policy says it “does not support” a tool used by many educators to evaluate whether a student used generative AI to complete an assignment.
Generative artificial intelligence is a tool that creates something based on an input, using years of information learned over the Internet to almost instantaneously respond to any prompt. Students can use it to help them research a paper or to write the paper entirely.
Many teachers have taken to using detectors that claim to be able to estimate how much of an assignment was completed by AI. Those tools have been discouraged by many experts, including the North Carolina Department of Public Instruction, because they are error-prone, producing both false positives and false negatives.
That’s a problem, a Green Hope High School student said, resulting in her receiving a zero on an essay she wrote herself.
This spring, Eleanor De Coster Canina called on the school board to prohibit the use of detectors.
District officials say internal district guidance for employees discourages the use of AI detectors. WRAL News requested that guidance last month but had not received it as of Tuesday afternoon.
Without public guidance, a student may not know that their teacher was discouraged from using an AI detector.
Using AI ethically
The draft policy calls unauthorized or undisclosed use of AI tools to complete schoolwork “academic misuse.” That would be addressed via the school board’s honor code and code of conduct policies.
The honor code policy addressed cheating, plagiarism, falsification and deceit. Under the honor code, a student could be required to redo an assignment or be disciplined more broadly, such as affecting a student’s eligibility for certain activities.
The draft before the school board calls for a human-centered approach to AI use in schools, including human oversight of any AI use.
Any use of AI for “official purposes” by staff or students would need to be disclosed and explained.
Staff use should promote engagement in authentic learning, the draft reads.
Students should do the same and avoid overreliance on AI. Students would only be able to use AI if authorized by their teacher and would need to explain it.
While the draft policy outlines rules, it doesn’t outline how the district would ensure compliance. Enforcement is often not included in policies but rather in internal procedures documents that can be more quickly edited because they don’t require school board approval.
Many students have told WRAL News that their classmates use AI to cut corners on their schoolwork. Many teachers, fearing that students will use AI and fail to learn new material, often avoid assignments that use technology at all.
DPI’s AI guidance, first issued in January 2024, includes an appendix of sample assignments that would use AI but require student actions intended to keep them engaged rather than simply going back to AI to complete the assignment.
For math lessons, DPI suggests students — rather than, say, solve for a variable in a typical algebraic equation — should use AI to create models for real-life scenarios to determine possible choices and evaluate the quality of those choices.
In a social studies class, students could use AI to draft arguments in a modern policy debate in 18th-century style, then compare rhetorical techniques from different eras. Or they could use AI to write three explanations for why the American Revolution occurred and evaluate each one against primary documents.
