AI and the Human Work of Teaching

Over the last few years, the world has seen an acceleration of what we call “artificial intelligence (AI),” a catch-all term referring to technologies that seem to have some pseudo-sentient element to it. This new wave of innovations includes generating long-form essays with just a few prompts to re-interpolating portraits into works of art from familiar artists. As a vocal set of artists, technologists, and intellectuals struggle with what these technologies mean for the future of their work, developers have sought to integrate AI into multiple facets of our on- and offline lives. These tools offer a premise: why spend hours on developing an artifact that AI can do in minutes? Much attention has been paid to the economic, environmental, and moral costs of such a proposition, though not enough to change the consumption of the general populace.

In education spaces, the amplification of AI has both terrified and emboldened educators – and by educator, I mean child-facing adults, inclusive of paraprofessionals, social workers, etc. – across the board. On social media, a plethora of educators are voicing either their displeasure or their support for using AI. On the one end, educators have strong arguments against students using AI for writing essays and solving math problems, arguing that using these tools ultimately dulls students’ critical thinking and ability to generate meaning from their assignments. To their credit, both the National Education Association and the American Federation of Teachers have sought to influence the ways AI affects teaching and learning. Last year, the NEA convened a nationwide panel of educators to create a report on the present and future role of AI in classrooms, including opportunities and harms. More recently, the AFT developed a partnership with Microsoft, OpenAI, Anthropic, and the United Federation of Teachers to create a “National Academy of AI Instruction,” a multimillion dollar effort to provide AI training and curriculum to K-12 educators.

On the other hand, districts have spent millions, if not billions, on AI-backed tools that write lesson plans, grade homework, and generate reports about students, teachers, and school performance. In other words, not only will AI tools have an effect on how we view learning, it necessarily reaches how we understand teaching. Recently, Microsoft announced a $4 billion donation to K-12 schools, colleges, and nonprofits as part of its “Microsoft Elevate” initiative. This follows a multi-tiered effort from big corporations to invest in AI in the classroom following President Trump’s executive order to foster more innovation in AI. The tech industry seems to push beyond school-level discussions about uses and misuses of AI towards large scale investments, but whether they implement input from community-level practitioners remains to be seen.

Teaching and learning as labor has not been examined enough from this lens. Teacher work has come under scrutiny through generative AI (GenAI) tools. A plethora of companies have done a great job of laying out the benefits of using GenAI to simplify the ardor of teaching. Some prognosticate that teachers would no longer be necessary in the near future. Given the already daunting teacher shortages across the country, negative attitudes about labor on the part of some district and federal leadership, and the acceleration of these tools, one might agree. (Education International, a global teachers’ union, is much more skeptical.) If, in the eyes of AI evangelists, we believe that teaching is just delivery of knowledge, what is the difference between an algorithm and a teacher? In the past, the pushback has prompted salespeople to treat teachers as “facilitators on the side.” In this rendition of teaching, educators “get out of the way” of students and their individual journey through the tool.

Holding for this framework, teaching as an occupation has already been deprofessionalized. Ironically, efforts to professionalize teachers in the way of standards and assessments have inversely stripped levels of autonomy and agency necessary for this authentic work. Simple notions of teaching as delivery belie the human-centered work of teaching i.e. the extent to which humans think about how younger humans learn through the meaning they make about the world. While researchers over decades have studied notions of professionalism among multiple professions, few, if any, ever ask the professionals themselves about their work and why it matters.

As such, it behooves us to rethink the question of AI and education. Instead of asking “How can we get more teachers to use our tools?”, we may ask “What parts of a teacher’s workload is made simpler using our tools?” While the former assumes teachers as simple recipients of advanced technologies, the latter affirms teachers’ responsibility to the trajectory of student learning in their classrooms. After all, even with the heightened use of AI, teachers report that AI is doing more harm than good in education. In surveying the research, a few things come to light:

1. Teachers have pushed back against the enormous amounts of paperwork and documentation that seems to sit outside of the periphery of their core work.
2. Teachers in the United States collectively have some of the highest rates of student-to-teacher facetime with seemingly few rewards. By comparison, other “high-performing” countries have significantly less time in front of students and more time to plan individually and collectively with colleagues.
3. Teachers are increasingly using GenAI to simplify the “tedium” of their work. It’s much simpler to plan lessons, create assessments, and assign class and homework through an advanced mechanism that can also pull together standards from prompts.
4. Teachers had already been using adaptive, data-informed technologies such as individualized assessment generators, learning management systems, and scripted lessons – willingly or otherwise – for at least the last 20 years.
5. Teacher morale sits in the backdrop of this conversation given the problems with teacher attrition.

Taken all together, this may seem like a perfect opportunity for education technologists to create more GenAI tools if they haven’t done so already. However, not only do the majority of students still want humans as teachers to teach them, schools don’t think teachers should be replaced with GenAI. In future essays, it’ll be important to go deeper into understanding the actual work of teaching and learning and where AI seems helpful or not.