The Human Judgment Doctrine: How Illinois’ Teacher Evaluation Ban, Maryland’s Training Mandate, and Idaho’s Anti-Replacement Law Are Redrawing the Line Around AI in the Teacher Classroom

The Human Judgment Doctrine: How Illinois’ Teacher Evaluation Ban, Maryland’s Training Mandate, and Idaho’s Anti-Replacement Law Are Redrawing the Line Around AI in the Teacher Classroom

As 134 bills flood state legislatures, a new consensus is forming: AI may assist teachers, but it cannot grade them, replace them, or leave them to figure it out alone.


The Policy Inflection Point

State lawmakers have introduced more than 134 bills across 31 states in 2026 addressing artificial intelligence in education, according to tracking by the government relations firm MultiState[1]. The legislation focuses on data privacy, classroom use restrictions, curriculum integration, and, critically for educators, professional development and oversight requirements. While much public attention has focused on whether students should use AI to write essays, a quieter and arguably more consequential debate is unfolding around the teacher’s side of the desk: Should AI be allowed to evaluate, score, or replace educators? And who is responsible for training teachers to use these tools at all?

The answer emerging from statehouses is a three-part doctrine. AI may support teachers administratively, but human judgment must control evaluation and high-stakes decisions. Teachers must receive mandated professional development, not ad hoc tutorials. And no AI system may replace a human teacher in the classroom. These principles are coalescing across red and blue states alike, reflecting a bipartisan consensus that AI in education is not a partisan issue but a professional one[1].


The Illinois Teacher Evaluation Ban

On July 10, 2026, Illinois Governor JB Pritzker signed Senate Bill 2909 into law, making Illinois the first state to expressly prohibit the use of AI tools in teacher evaluations[2]. The law, spearheaded by State Senator Christopher Belt (D-Swansea), amends the Evaluation of Certified Employees Article of the Illinois School Code to bar evaluators from using AI to assign numerical scores or qualitative ratings for any component of a teacher’s evaluation[3].

Senator Belt’s rationale was straightforward. “I believe that our teachers should be judged based on actual observations and professional judgement, not by AI software,” he said. “Our educators deserve a transparent and fair evaluation process that demonstrates their actual work in the classroom and protects their privacy”[2].

What the Law Prohibits

The legislation draws a precise line between evaluative judgment and administrative support:

  • Prohibited: AI tools may not assign numerical scores, generate qualitative ratings, or perform any evaluation task requiring professional judgment. This covers scoring classroom observations, assessing instructional effectiveness, and generating performance review summaries that constitute qualitative ratings[3].
  • Prohibited: Teachers may not use AI tools to generate evidence of their professional practice for evaluation purposes. The restriction runs in both directions[3].
  • Permitted: AI may be used for administrative support tasks, such as transcribing observation notes, organizing data, or scheduling evaluations, provided the tool does not influence the evaluative outcome[3].
  • Required disclosure: If an evaluator or teacher uses any AI tool for permitted administrative purposes, they must disclose the tool’s name and specific purpose to the other party. This transparency requirement applies to both sides of the evaluation process[3].

The core logic here is that teaching is a deeply contextual, qualitative practice. Classroom management, student engagement, instructional adaptability, and the creation of a positive learning environment are not dimensions that can be ethically reduced to an algorithmic score. The law’s architects recognized that AI might eventually become reliable enough for this purpose, but the technology is not there yet. By codifying the human-judgment requirement now, Illinois is establishing a precedent that other states are watching closely[3].

The law also carries a practical compliance burden. School districts must review and revise their teacher evaluation plans to incorporate the new prohibitions and disclosure rules. Collective bargaining agreements may need amendment. Evaluators and teachers will need training on what constitutes a prohibited use versus a permissible administrative use. The Illinois State Board of Education is expected to provide guidance, but the bill does not create a new oversight body. It relies on existing district-level compliance mechanisms and the established grievance procedures for alleged violations[3].


The Professional Development Mandate Wave

While Illinois drew a hard line on evaluation, other states are addressing a more basic problem: the vast majority of teachers have received little to no training on AI. A 2024-25 survey by the Center for Democracy & Technology found that 85 percent of teachers reported using AI in their classroom, but only about half said they received any training or information about AI from someone at their school, and even fewer received training on the risks of AI use[4]. Maryland State Senator Katie Fry Hester described the problem bluntly: “What we heard repeatedly is that the teachers were feeling like they had to navigate artificial intelligence entirely on their own”[5].

Maryland’s Artificial Intelligence Ready Schools Act

Maryland’s SB 720, signed into law in May 2026, takes the most comprehensive approach to teacher professional development of any state this session[6]. The law:

  • Requires compensated statewide professional development through a train-the-trainer model, meaning the state funds the training rather than leaving districts to absorb the cost[6].
  • Mandates AI coordinators in each local school system, creating a designated human point of contact for AI policy implementation[6].
  • Establishes a statewide AI Education Collaborative to study AI in K-12 education and share best practices across districts[6].
  • Integrates AI into workforce standards and educator training, treating AI literacy as a core professional competency rather than an optional add-on[1].
  • Requires university-supported certification of compliant AI tools, creating a procurement gatekeeper that prevents districts from buying unvetted technology[1].

The train-the-trainer model is particularly significant because it solves a scale problem. Maryland cannot send every teacher to a multi-day AI workshop. But it can train a cadre of experienced educators who then train their colleagues, creating a cascade of professional development that reaches every school system without requiring a massive new bureaucracy. The compensation requirement ensures that this training happens on paid time, not as an unfunded mandate that teachers absorb on evenings and weekends[6].

The Federal Executive Order

State laws are building on a federal framework established in April 2025, when President Trump signed Executive Order 14277, “Advancing Artificial Intelligence Education for American Youth.” The order declares it the policy of the United States to promote AI literacy by “providing comprehensive AI training for educators” and directs multiple federal agencies to act[7].

Specifically, the order required the Secretary of Education to issue guidance within 90 days on the use of formula and discretionary grant funds to improve education outcomes using AI, including AI-based instructional resources, high-impact tutoring, and college and career pathway exploration. It also directed the NSF and the Secretary of Agriculture to utilize existing programs to create teacher training opportunities that help educators integrate AI-based tools into classrooms[7]. The Department of Education subsequently issued guidance making AI in education a grantmaking priority, and the White House launched a pledge that now includes commitments from over 200 tech companies and organizations, including Google, OpenAI, Microsoft, NVIDIA, and Salesforce[4].

The American Federation of Teachers responded by launching the National Academy for AI Instruction in partnership with Microsoft, OpenAI, Anthropic, and the United Federation of Teachers, a $23 million initiative to train educators on AI integration[4]. This union-led, industry-funded model represents a parallel track to state mandates: voluntary, well-resourced training that reaches teachers through their professional associations rather than through legislative compulsion.

Tennessee’s Failed Certification Mandate

Not every professional development proposal has succeeded. Tennessee’s SB 677, introduced during the 2025 session, would have required the state Department of Education to provide a professional development course on the use of AI in the classroom and mandated that teachers complete such a course within two years of becoming certified[4]. The bill did not pass, but it illustrates the more aggressive end of the mandate spectrum: tying AI training directly to teacher certification, the legal permission slip to teach. Had it succeeded, Tennessee would have created a structural gatekeeper ensuring no new teacher enters a classroom without AI training, a model that other states may revisit in future sessions.


The Anti-Replacement Principle

Beyond evaluation and training, a third principle is gaining statutory force: AI must not replace human teachers. Idaho’s SB 1227, signed into law in March 2026, establishes a statewide framework for generative AI in K-12 schools that explicitly ensures no AI “replaces or eliminates a human teacher”[5]. The law also requires state standards for AI literacy and educator training, mandates local school district policies for AI usage, and sets data privacy requirements for AI tools in schools[1].

Oklahoma’s SB 1734, the Responsible Technology in Schools Act, enacted in May 2026, takes a complementary approach. It allows AI in schools only under educator supervision with human review, bans its use for high-stakes decisions including grading, discipline, and placement, and requires every district to adopt a written AI policy before the 2027-28 school year. Parents retain the right to opt their children out of student-facing AI tools without academic penalty[6]. The principle is consistent across both laws: AI is a tool that teachers direct, not a system that directs teachers.

These anti-replacement provisions matter because they address a fear that educators have voiced since ChatGPT’s rollout in 2022: that budget-strapped districts might see AI as a cost-saving substitute for human instruction. By embedding the human-teacher requirement in statute, Idaho and Oklahoma are making it legally impossible for a district to replace a teacher with an AI system, even if the economics make it tempting. The laws turn a professional norm into a legal right[5].


The SREB Professional Learning Framework

Beneath the legislative activity, education organizations are developing the practical frameworks that states and districts will need to implement these mandates. In March 2026, the Southern Regional Education Board published “Considerations for Designing AI Professional Learning for Educators,” a four-phase framework developed by the SREB Commission on AI in Education, chaired by South Carolina Governor Henry McMaster and co-chaired by Marshall University President Brad D. Smith[8].

The framework recommends a phased approach that builds in the complexity of AI use over time:

  • Phase 1, Building Awareness: Educators explore AI through guided activities, using it to generate lesson ideas, translate materials, and analyze assessment data. The focus is on adult use first, allowing schools to procure compliant tools before introducing students to AI[8].
  • Phase 2, Exploration: Educators develop a foundational understanding of what AI is and is not, common terms, and prompting strategies. This phase includes adopting procurement checklists that examine AI tools for capabilities, human oversight, bias mitigation, evidence of impact, security, and data use[8].
  • Phase 3, Application: The focus shifts from adult-only use to allowing students to use AI within the learning environment, with teachers designing cognitively demanding tasks that use AI as part of the learning process rather than circumventing it[8].
  • Phase 4, Iteration: Schools and districts expand their AI toolboxes while cultivating a culture of curiosity, creativity, and responsible technology use, integrating AI career readiness with professional learning outcomes[8].

The framework’s core insight is that professional development cannot be a single workshop. AI is evolving too quickly for one-and-done training. By structuring professional learning as a multi-year progression from awareness to fluency, SREB is acknowledging that the technology teachers will use in 2028 will be different from what they use today. The framework also emphasizes that leadership must attend the same sessions as teachers, so that administrators understand what they are asking their staff to adopt[8].


The Skeptical Counterpoint

Not everyone agrees that the current wave of legislation is well-calibrated. Justin Reich, an associate professor of digital media at MIT, argues that states are trying to regulate classroom AI at a moment when the field is still so unstable that “writing a guide for AI in 2026 is like writing a guide for aviation in 1905” before airlines, airports, or even commercial flight existed. “If you were to take any of the AI literacy documents, AI readiness documents, even the moratorium documents, and put them against a checklist,” Reich said, “there would be a lot of boxes in the ‘we’re making this up’ column and not a lot in the ‘we have evidence’ column”[5].

Reich’s caution is not an argument against regulation but a warning against false confidence. The risk is that states codify specific AI tools or practices into law, only to find that those tools are obsolete within two years. Noelle Ellerson Ng of the School Superintendents Association makes a related point: schools should not cover AI in a separate policy but as part of a broader technology policy. “I don’t have a calculator policy,” she said. “Why would I have an AI policy? I have a technology policy”[5].

There is also evidence that the current approach may not be reaching teachers fast enough. The Center for Democracy & Technology’s survey found that half of students said using AI in class made them feel less connected to their teachers, and 70 percent of teachers said they were concerned that students’ use of AI was preventing them from learning important skills[5]. Only 11 percent of teachers said they had been trained on how to respond if they suspected a student’s AI use was detrimental to their well-being[4]. These numbers suggest that even as states pass training mandates, the gap between policy and classroom reality remains wide.


Where This Is Heading

The 2026 legislative session is producing a coherent set of teacher-facing AI policies that share three structural commitments: human judgment controls evaluation, professional development is mandated and funded, and AI cannot replace a teacher. The specific mechanisms vary, from Illinois’ outright prohibition on AI scoring to Maryland’s train-the-trainer model to Idaho’s anti-replacement statute. But the underlying principle is consistent across party lines and geographic regions.

Looking ahead, several unresolved questions will shape the next phase of policy. First, whether the human-judgment requirement in teacher evaluations will extend to other high-stakes domains, such as student grading or disciplinary decisions. Oklahoma already bars AI from high-stakes student decisions[6], but most states have not addressed this. Second, whether the professional development mandates will be adequately funded, or whether they will become unfunded mandates that districts struggle to implement. Maryland’s compensated model is the exception, not the rule. And third, whether the anti-replacement provisions will hold as AI systems become more capable of delivering instruction independently, creating economic pressure that statutes alone may not withstand[5].

For now, the message from state legislatures is clear: AI in the teacher classroom is welcome, but on terms that educators, not algorithms, set[1].


References

[1] [SECONDARY] Aaron, Izzy. “How States Are Regulating AI in Education this Legislative Session.” MultiState Insider. April 9, 2026. https://www.multistate.us/insider/2026/4/9/how-states-are-regulating-ai-in-education-this-legislative-session

[2] [PRIMARY] Illinois Senate Democratic Caucus. “Illinois bans AI use for teacher evaluations.” July 10, 2026. https://www.illinoissenatedemocrats.com/caucus-news/20-sen-christopher-belt-news/7027-illinois-bans-ai-use-for-teacher-evaluations

[3] [PRIMARY] Regulations.ai. “Illinois AI Teacher Evaluation Ban, SB 2909.” 2026. https://regulations.ai/regulations/RAI-US-IL-SB29090-2026

[4] [ADVOCACY] Center for Democracy & Technology. “States Focused on Responsible Use of AI in Education during the 2025 Legislative Session.” 2025. https://cdt.org/insights/states-focused-on-responsible-use-of-ai-in-education-during-the-2025-legislative-session/

[5] [SECONDARY] Sequeira, Robbie. “As AI use in schools grows, lawmakers and districts scramble to set up guardrails.” Stateline. June 10, 2026. https://stateline.org/2026/06/10/as-ai-use-in-schools-grows-lawmakers-and-districts-scramble-to-set-up-guardrails/

[6] [SECONDARY] DiMarco, Bella. “Legislative Tracker: 2026 State AI in Education Bills.” FutureEd. March 2, 2026 (updated July 13, 2026). https://www.future-ed.org/legislative-tracker-2026-state-ai-in-education-bills/

[7] [PRIMARY] The White House. “Advancing Artificial Intelligence Education for American Youth.” Executive Order 14277. April 23, 2025. https://www.whitehouse.gov/presidential-actions/2025/04/advancing-artificial-intelligence-education-for-american-youth/

[8] [SECONDARY] Eaves, Leslie, and Jason Adair. “Considerations for Designing AI Professional Learning for Educators.” Southern Regional Education Board. March 2026. https://www.sreb.org/sites/main/files/file-attachments/2026-designing-ai-pl-for-educators.pdf


AI automated and generated content. Human reviewed.

The Weekly AI Policy Brief

AI policy moves fast. Get the week’s developments, new trackers, and podcast episodes in one short email — free, ad-free, no spam.

Comments

Leave a Reply

Your email address will not be published. Required fields are marked *