New and advancing artificial intelligence tools can be effective accommodations for students with disabilities to support their participation in the general education curriculum. But accommodations are not instruction. IEP teams will need to carefully consider when such tools are appropriate and when specialized instruction in underlying skills is warranted. Teams should also take care not to overlook learning gaps when students independently employ AI tools as compensatory strategies to mask lagging skills. A recent Tennessee case highlights this issue and reminds teams that grades may not provide a clear signal of a student’s progress.

In W.A. v. Clarksville-Montgomery County School System, the parents of a high school student with dyslexia challenged the school’s program for failing to provide goals and instruction in foundational reading skills. The school, however, argued that the student was earning passing grades, including on reading and writing assignments. The court found that, although passing grades are significant evidence in favor of finding the student received a free and appropriate public education (FAPE), in this case, the student’s grades did not sufficiently or accurately demonstrate appropriate progress. Technology allowed the student “to bypass, rather than address, his lack of basic reading skills due to his dyslexia.”

The evidence showed that the student earned passing grades by using “complex technological workarounds that allowed him to bypass the need for reading and writing skills.” The student’s system involved multiple steps and tools:

1. Use speech-to-text to enter a topic or prompt into a Word document.
2. Paste the language into Chat GPT to generate an answer or essay.
3. Paste the response into Word.
4. Use Grammarly to adjust the style and make grammatical corrections.
5. Use read-aloud software to listen to the revised response.
6. Make changes where he was not satisfied (presumably using speech-to-text again).
7. Use Grammarly to fix any errors he introduced.
8. Listen again using the read-aloud software.

Impressive, right? Students with dyslexia finding their own hacks and strategies to mask their reading struggles is not a new phenomenon, but artificial intelligence provides new and powerful ways to do so. The question for IEP teams is: When is AI an effective tool to enable the student to participate and progress in the general education curriculum and prepare for postsecondary life, and when is it an inappropriate shortcut that deprives the student of learning and practicing fundamental skills?

This too is not a new conundrum. For years, IEP teams have similarly wrestled with the question of when it is appropriate to provide a student specialized instruction to learn math calculations and when it is appropriate to provide the student a calculator as an accommodation to support math problem solving. The answer, as always, needs to be individualized to the student, considering factors such as age, the impact of the disability, prior interventions and progress, and long-term goals. Additionally, the answer does not have to be either-or, some students will continue to work on math calculation goals while also using a calculator for other assignments.

Similarly, the tools used by the student in this case may be appropriate accommodations for some assignments. But the court found that an appropriate IEP needed to provide dyslexia-focused instruction to improve the student’s reading skills. Citing Endrew F., the court instructed that “simply finding some combination of accommodations and lenience that allows a child to graduate is not necessarily enough.”

Especially as students and parents become more adept at using new technologies, IEP teams will increasingly need to address when AI tools are an appropriate accommodation, when further skill development is appropriate, and when both may be warranted. Rather than condemning tools like Chat GPT and Grammarly as cheating, IEP teams should consider whether a student is using the tools to make up for gaps in skills that the IEP should address directly, and whether there may be appropriate use cases. As with all IEP team decisions, open discussions, data-based reasons, and clear documentation will put the team in a strong position in the event of a later challenge.

Please reach out to our Special Education team with questions about AI and special education.