Picture this: A colleague spends Sunday night feeding her lesson outline into an AI tool, gets back a polished set of discussion questions, and walks into class feeling prepared. When a student pushes back—“I’m not sure this question even fits the reading”—she has to respond in real time, drawing on fifteen years of watching students struggle with exactly that concept. No AI gave her that. That’s the point.
Over the last few years, tools like ChatGPT and other generative AI programs have moved from novelty to near-necessity in our teaching lives. Commentators have been asking whether AI can replace teachers for years, and they consistently point back to the irreplaceable human dimensions of mentoring, judgment, and care. While adaptive and AI-driven systems can personalize content, they cannot replicate the relationship between a teacher and a student. That relationship is exactly what is now being protected in California law.
In California’s community colleges, artificial intelligence no longer lives only in the realm of “emerging technology”; it now intersects with something else that matters deeply to us: state legislation. Recent action in Sacramento clarifies that while AI can support our work, it cannot replace us as instructors of record. In practical terms, it tells us what many of us already know from experience: AI can help, but the teacher stays at the center.
What does that look like in a real classroom? Below are four moves that can help you integrate AI in ways that are both legally sound and pedagogically grounded.
Move 1: Use AI as a Research Assistant, Not a Replacement
California’s Assembly Bill 2370, signed into law in 2024, adds Section 87709 to the Education Code for community colleges. The law states that the instructor of record for a course must meet the minimum qualifications established by the Board of Governors of the California Community Colleges and explicitly prohibits using AI to replace faculty in providing academic instruction or regular interaction with students. AI can be used to support course development, assessment, and tutoring, but it cannot stand in for a qualified human instructor.
One way to operationalize this is to treat AI like a research assistant rather than a co-teacher. Let it handle the legwork, not the judgment. For example, you might:
- Ask an AI tool to generate ten multiple-choice questions from a reading, then select and revise the three that target your outcomes.
- Draft a rubric with AI in two minutes, then spend ten minutes adjusting language so it reflects what you value in student writing.
- Use AI to brainstorm possible case studies or scenarios, then choose the ones that best fit your students lived experiences.
In each case, you are still doing the core intellectual work: deciding what matters, how it should be assessed, and how it fits your course. If AI is making the pedagogical decisions, you have moved beyond assistance into replacement—and that is precisely what the law is designed to prevent.
You can make this explicit for students. A brief note in your syllabus might read, “I may use AI tools to help generate ideas for questions or assignments, but I always review, revise, and decide what we use in this class. Your instructor—not AI—is responsible for your instruction and grades.” That one sentence echoes the spirit of AB 2370 and reassures students that a human is in charge.
Move 2: Protect Live Teaching as Human Space
The statute’s language around “academic instruction and regular interaction with students” can guide how we think about AI during live class time. Put simply: planning can be AI-assisted, but the room (physical or virtual) belongs to you and your students.
Consider a typical 75-minute class. Before class, you might ask AI to:
- Suggest three ways to introduce a tricky concept.
- Generate a short scenario you can use for a quick-write.
- Propose discussion prompts at different difficulty levels.
Once class begins, put the AI away. Use your expertise to read the room: noticing who is confused, who is disengaged, and who is ready to move on. When a discussion veers in an unexpected direction, you draw on your experience, not a pre-scripted AI plan. When a student’s life circumstances surface in office hours, you respond as a mentor, not an algorithm.
This vision aligns with the “always center educators” model of human-in-the-loop AI (U.S. Department of Education, Office of Educational Technology, 2023). The report emphasizes that AI should be designed and implemented in ways that center educators in three areas: preparing for and reflecting on teaching, engaging in the act of teaching, and designing and evaluating tools.
Figure 1: Three ways to center educators as we conceptualize human-in-the-loop AI.
In this framework, AI can help you analyze patterns in student work, generate draft materials, or suggest alternative explanations, but you remain the one who interprets the data, leads the instruction, and decides which tools support your course goals. You might even narrate this for students: “I sometimes use AI to brainstorm ideas before class, but I don’t use it to run our meetings. What happens in this room is based on you and what I see you doing.” That statement both reflects the law’s emphasis on real human interaction and gives students a language for thinking critically about AI in their own learning.
Move 3: Evaluate AI the Way You Evaluate Textbooks
AB 2370 assumes that faculty will use professional judgment when incorporating AI. One practical way to exercise that judgment is to evaluate AI tools the way you evaluate textbooks or other course materials.
When you consider adopting a new textbook, you probably skim the content, check for alignment with your learning outcomes, and look for bias or gaps. Apply the same process to AI:
- Run your own prompts first. Before recommending a tool to students, ask it to perform the tasks you plan to assign (summarizing an article, explaining a concept, modeling a paragraph).
- Check for accuracy and bias. Compare its responses with your disciplinary knowledge. Note where it oversimplifies, where it leaves out key voices, or where it reflects dated or narrow perspectives.
- Align with outcomes and policy. Ask, “Does using this tool help students practice the skills this course is designed to build, or does it short-circuit the learning process?”
You might build a short, low-stakes activity around this evaluation. For example, in a writing class, ask students to generate a draft paragraph with AI and then critique it using your rubric: “What does the AI do well? Where does it fall short of our standards?” This not only demystifies AI but also reinforces your expectations.
In your syllabus or LMS, you can include a brief “AI Tools” section outlining which tools are allowed, for what purposes, and under what conditions. Tying this back to the legislative context—explaining that state law requires a qualified instructor, not AI, to provide instruction and interaction—can clarify why these boundaries exist.
Move 4: Turn Policy into Student-Facing Conversations
Legislation often feels distant from our day-to-day teaching, but AI policy is already shaping student expectations. Some students assume AI is off-limits; others assume it is required. The law gives you a reason to have explicit, human-centered conversations about it.
You might open a class session with a prompt like:
- “In what ways do you currently use AI for school, work, or life? Where do you draw the line?”
- “What do you think a ‘good teacher’ does that AI cannot do?”
- “Given that state law says AI cannot replace your instructor, how should we use it—if at all—in this course?”
Have students respond in a quick-write, then discuss in pairs or small groups before a whole-class debrief. As they talk, you can connect their ideas to your course policies and to the broader context: national groups tracking AI-related legislation have noted that the number of AI bills introduced in state legislatures more than doubled in a single year, and California is among the states explicitly grappling with AI in education. Framing your classroom choices within that larger trend can help students see that your policies are not arbitrary—they are part of a wider effort to use AI ethically and responsibly.
You can revisit this conversation mid-semester. Ask students how their AI use has changed and what they are learning about when AI helps and when it hurts. These reflective checkpoints reinforce the message that they are not passive recipients of technology, but active participants in shaping how it’s used.
California’s AI law gives us something rare in higher education: a clear statement that our presence and qualifications matter in an age of rapid technological change. It does not forbid AI, nor does it require us to use it. Instead, it invites us to experiment—carefully—while remaining firmly in the driver’s seat.
When we use AI as an assistant rather than a replacement, protect live teaching as human space, evaluate tools with the same care we use for textbooks, and talk openly with students about policy and practice, we do more than comply with legislation. We model the kind of critical, ethical engagement with technology that our students will need long after they leave our classrooms. And that is something no AI can do for us.
Note: This article was the result of a collaboration between the human author and AI programs (ChatGPT and Claude) used for idea generation and drafting, with all final decisions and revisions made by the human author.
David E. Balch, PhD, is a professor at Rio Hondo College and has published articles in the areas of ethics, humor, distance education, and AI.
Robert Blanck, MA, has 35 years’ experience in teaching and administration within the fields of education and business. Blanck earned the Lean Six Sigma Black Belt designation for industrial quality control.
References
Bill Text – AB 2370. (2024). Community colleges: faculty: instructor of record: qualifications. California Legislature.
Dusseault, B., & Lee, J. (2023). AI is already disrupting education, but only 13 states are offering guidance for schools. Center on Reinventing Public Education.
Frazzini, K. (2024). Volume of AI bills rises, even as use of systems evolves. National Conference of State Legislatures.
Jones, B. M. (2023). How educators can use generative AI to promote student innovation. Forbes.
Kolchenko, V. (2018). Can modern AI replace teachers? Not so fast! HAPS Educator, 249–251.
Lake, R. (2024). AI is already disrupting education, but only 13 states are offering guidance for schools. Center on Reinventing Public Education.
Mathews, A. (2024). FACCC-sponsored AB 2370 is approved by the legislature. Faculty Association of California Community Colleges.
U.S. Department of Education, Office of Educational Technology. (2023). Artificial intelligence and the future of teaching and learning: Insights and recommendations.