Maryland School AI Policies Arrive This Fall. One Critical Gap Remains.

What Maryland’s Law Requires and When

Maryland’s AI Ready Schools Act (Senate Bill 720, signed as Chapter 634) tasks every school district with adopting an AI policy before fall 2026. The Maryland State Department of Education presented its guidance to the State Board in February and held the formal release while ensuring legal alignment with the new legislation. Under the framework, each of Maryland’s 24 local school systems must address eight elements: data privacy, technology bias, approval criteria for AI tools, human-centered instruction, equity protections, and professional development for educators, among others. The state will publish online modules to help teachers level-set their understanding of AI tools as the technology evolves.

Senator Katie Fry Hester, a sponsor of the legislation, framed the urgency in workforce terms. “Whether a student wants to be a nurse or a teacher or a mechanic or an engineer or if they want to start a business, they’re going to encounter and have to use artificial intelligence,” she said. “The students need to know how to use these tools, and they need to understand the fundamentals of how they work.” The law also establishes a Maryland AI Education Collaborative drawing on educators, administrators, parents, and students to study AI use in schools and report findings annually.

Legal Compliance and Pedagogical Effectiveness Are Two Different Things

The mandate’s legal requirements are genuine improvements. Requiring IDEA and Section 504 compliance means school districts are barred from deploying AI tools that lack accessibility features. Monitoring for bias and disparate impact means the equity concern is formally on the record. These are not nothing. But legal compliance with disability law is not the same as demonstrated effectiveness for students whose brains process information differently. Those are two separate tests, and Maryland’s framework only runs one of them.

A 2026 systematic review published in Brain Sciences analyzed 11 studies covering more than 3,000 students with learning disabilities and found meaningful outcomes from AI interventions, but only when those tools were embedded within a structured pedagogical framework, not deployed as standalone programs. The same review identified five significant challenges limiting the field, and bias and validity in assessments was among them: AI systems built primarily on typical-learner data risk generating inaccurate progress pictures for children with auditory processing differences, slow processing speed, or working memory challenges. That gap exists even when legal standards are fully met.

This is what Maryland’s eight-point checklist does not close. A district passes every point on the framework and still adopts a tool designed for average learners, tested on average learners, with no published data on how it performs for a child who struggles with auditory processing or visual tracking. Reading is not one skill. Math is not one skill. Attention is not one skill. An AI tool built around a single-system model of learning faces the same ceiling that phonics-only instruction faces: it addresses part of the problem and misses the rest. One-tool-fits-all is the same old failure wearing a new interface.

What Parents in Maryland (and Every Other State) Should Do Now

For parents in Maryland, the fall policy deadline is an opening, not a finish line. When your district publishes its AI policy, that document tells you which tools are approved. It does not tell you whether any of those tools were evaluated with students who have reading, attention, or processing differences.

Two questions are worth bringing to your school board’s AI policy meeting. First: Was this tool evaluated with students who have learning or processing differences, and what did those results show? Second: Does the tool adapt to processing speed and working memory differences, or does it adapt only to content difficulty? The difference is not small. A child with slow processing speed needs more time and repetition at the right level, not harder content. A child with auditory processing differences needs a different instructional format, not a longer reading passage. A tool that adapts only to content will miss both of them.

Maryland’s law is a real step. Other states will follow. The pattern the system runs well is setting legal minimums and calling it progress. Parents who know what legal minimums cover, and what they do not, are the ones who know what to ask next.