What does the research say about AI and student learning?

The clearest evidence comes from a field experiment involving nearly 1,000 high school math students. Bastani et al. (2025, PNAS) found that access to a standard ChatGPT-style tool increased performance on practice problems by 48 percent. But when access was removed, those students scored 17 percent lower on the exam than students who never had it. A tutor version designed to give hints rather than answers increased practice performance by 127 percent without the exam penalty.

The difference was not the technology. It was who did the thinking.

Survey research points the same way. In a study of 666 participants, Gerlich (2025, Societies) found that frequent AI use was associated with lower critical-thinking scores, a link explained by cognitive offloading, which means handing mental work to the tool.

I see this in my own classroom. In my environmental economics course, about 18 percent of first-week reflections in 2022 listed the same combination of environmental problems that ChatGPT typically returns. By 2023, that share was 48 percent, and it was still 45 percent in 2025. I wrote about those three years of data in Teaching Economics in the Age of AI.

How I use AI as an economics professor

I use AI every week, for tasks where my judgment stays in charge.

Weekly reminder emails. I gave an AI tool my syllabus. Every Friday, it drafts a reminder email for students covering what is due that weekend, what is coming next week, and what to look forward to. I review the draft and then post it to our learning management system. The AI saves me time on formatting. I still decide what students need to hear.

Interactive worksheets. I also use AI to build interactive worksheets where students shift a market, adjust inputs, and see a concept respond, while they still do the reasoning. I collect them in Let's Do Classwork!

In both cases, AI handles the routine work, and the thinking stays with a person.

A simple test: who did the thinking?

Before you use AI on an assignment, ask one question: Is this the step I am supposed to learn?

• Use AI to explain a concept in a different way, quiz yourself, or check your reasoning after you have tried.

• Do not use AI to produce the answer to a problem you are expected to solve yourself.

Using AI to understand why a firm maximizes profit where marginal revenue equals marginal cost supports learning. Using AI to solve the assigned profit-maximization problem replaces the need to solve it. For a classroom example of this line in practice, see They All Got A's. None of Them Learned Anything.

Frequently asked questions

Does using ChatGPT make students worse at thinking?

Not by itself. Students lose skills when AI completes the reasoning a task was meant to build. In a field experiment with nearly 1,000 students, answer-giving AI lowered exam scores by 17 percent, while a hint-based tutor did not.

What is cognitive offloading?

Cognitive offloading means handing mental work to a tool. Some offloading is useful, like using a calculator. It becomes a problem when the offloaded step is the skill you are supposed to learn.

How can students use AI without becoming lazy thinkers?

Ask one question first: is this the step I am supposed to learn? If yes, attempt it yourself, then use AI to explain, quiz you, or check your work.

Should professors ban AI in their courses?

In my experience, redesigning assignments works better than banning AI. Students do the reasoning in class, and AI handles routine tasks or gives feedback.

Educators: What is one task you hand to AI, and one you never will? Share it in the comments.

Know someone confused about AI use in education? Share this post with them, whether a student, a parent, or a fellow educator. If it resonated, please like and restack it so it reaches more people asking the same question.

Read more on teaching and learning with AI:

• They All Got A's. None of Them Learned Anything.

• Teaching in the AI Age: Why Flipped Classrooms Are Working Better Than Ever

• Turning AI answers into a learning tool starts with intentional assignment redesign

Mahelet G. Fikru is a professor of economics at Missouri University of Science and Technology with 15 years of teaching and research experience.

Let’s Do Classwork! is a free collection of interactive economics exercises. I built them from the problem sets I use in my applied microeconomics courses at Missouri University of Science and Technology. Each lab lets students change inputs, see the result right away, and check their answers. The labs cover supply and demand, marginal analysis, elasticity, cost curves, game theory, market power, environmental economics, and U.S. pollution abatement cost data. Undergraduates, graduate students, and self-learners can use them to review a concept before an exam or to prepare for graduate coursework.

At a glance: 11 free interactive labs (seven microeconomics labs plus a four-part environmental economics lab)

· Levels: undergraduate, graduate, and general audience

· Created by Mahelet G. Fikru, Professor of Economics, Missouri S&T

· Last updated September 2026.

Find a lab by level

• Undergraduate review: The Efficiency Lab

• Graduate review: the Equilibrium, Nash, Elasticity, Cost, and Market Power Labs

• No economics background needed: The Workforce Lab

• Environmental and resource economics: The Environmental Economist Lab

1. The Equilibrium Lab: Supply and Demand, Profit Maximization, and Risk

This interactive exercise covers three fundamental economic concepts in one handout: shifting a market to see how supply and demand respond to changing conditions, chasing a profit-maximizing output level to find where a firm should produce, and pricing two risky investments to compare returns against risk. Click through each one, adjust the inputs, and see the concept play out in real time instead of just reading about it.

Concepts covered: market equilibrium, shifts in supply and demand, profit-maximizing output, expected return and risk.

Recommended for graduate students who need to review concepts in microeconomics.

2. The Nash Lab: Game Theory and the Principal–Agent Problem

This interactive exercise covers the agency problem between managers and owners — why managers acting on shareholders' behalf don't always choose what's best for the firm, and how strengthening monitoring changes the payoffs enough to shift the equilibrium outcome.

Concepts covered: Nash equilibrium, payoff matrices, the principal–agent problem, monitoring and incentives.

Recommended for graduate students who need to review concepts in game theory.

3. The Efficiency Lab: Marginal Analysis, Consumer and Producer Surplus, and Taxes

This interactive exercise covers the efficiency of markets — finding the optimal input level where marginal benefit meets marginal cost, deriving market equilibrium from supply and demand, and measuring the net benefits a market creates through consumer and producer surplus, including how a per-unit tax changes that split.

Concepts covered: marginal benefit and marginal cost, market equilibrium, consumer surplus, producer surplus, tax incidence.

Recommended for undergraduate students who need to review marginal analysis and market surplus.

4. The Workforce Lab: Data Center and Computing Jobs (NAICS 518)

An interactive look at who works in Sector 518 (computing infrastructure, data processing, hosting, and related services) and how that's shifted over time. Explore employment broken down by sex and by education level from 2010 to 2025, including a widening gender gap. Since data center jobs fall within this sector, tracking Sector 518 is one of the few ways to see AI-driven data center employment show up in the numbers, even though it isn't broken out on its own.

Concepts covered: employment trends, the gender gap in tech employment, education levels, how AI and data center jobs show up in industry data.

Data: U.S. Census Bureau, LEHD Quarterly Workforce Indicators (QWI).

Recommended for anyone with no formal training on economics or data analysis.

5. The Elasticity Lab: Price, Cross-Price, and Income Elasticity of Demand

This interactive exercise covers price, cross-price, and income elasticity of demand — calculating and interpreting each measure, along with how elasticity connects to total revenue (TR) and marginal revenue (MR).

Concepts covered: price elasticity, cross-price elasticity, income elasticity, total revenue, marginal revenue.

Recommended for graduate students in applied microeconomics who have a good command of analytical skills and a solid understanding of the demand-supply model.

6. The Cost Lab: Short-Run and Long-Run Cost Curves

This interactive exercise covers the structure of short-run production costs, deriving the relationships between total, fixed, and variable costs; using marginal cost to find how the cost of the next unit changes, and distinguishing accounting profit from economic profit through an opportunity-cost decision, including a look at how the long-run cost curve differs from the short-run one.

Concepts covered: total, fixed, and variable cost; marginal cost; opportunity cost; economic vs. accounting profit; long-run average cost.

Recommended for graduate students who need to review cost curve relationships and the accounting-vs-economic-profit distinction.

7. The Market Power Lab: From Perfect Competition to Monopoly

This interactive lab walks students from perfect competition to monopoly through guided stations. Students derive equilibria, fill in calculation tables, build an Excel model, and get instant feedback as they go.

Concepts covered: perfect competition, monopoly pricing, deadweight loss, building the model in Excel.

It is recommended for graduate applied microeconomics as a hands-on refresher on imperfect competition.

8. The Environmental Economist Lab

A. Externalities

A steel mill pollutes the air and never pays for the damage. Solve the competitive and efficient equilibria, add marginal damage cost to get marginal social cost, and measure the deadweight loss of a negative externality.

B. The Commons Lab

Public goods and common-pool resources side by side: a climate abatement game between two emitting blocs, a shared wetland buffer solved with the Samuelson condition, and a groundwater aquifer facing the tragedy of the commons. Shows why the free-rider problem and the tragedy of the commons come from the same root: non-excludability.

C. Nuclear Generation in the US

Who works in U.S. nuclear power plants (NAICS 221113), by state and by sex, quarter by quarter from 2007 to 2025. Data: U.S. Census Bureau, LEHD Quarterly Workforce Indicators.

D. Abatement Cost Atlas: U.S. Manufacturing Pollution Abatement Costs by State and Industry (2005)

How much did U.S. manufacturers spend to reduce pollution, and where did that money go? In 2005, plants spent about $20.7 billion on pollution abatement operating costs and $5.9 billion on abatement equipment. This interactive atlas turns the raw Excel tables from the Pollution Abatement Costs and Expenditures (PACE) survey into a map and filterable charts. Compare spending by state and manufacturing sector, and break it down by activity (treatment, prevention, recycling, disposal), by media (air, water, solid waste), and by cost category (labor, energy, materials, contract work, depreciation). Data: U.S. Census Bureau, PACE 2005 survey.

More coming soon

This is the start of a series. More classroom exercises will turn into these interactive pieces and land here as they’re made. Try them out, and share what’s confusing, what’s fun, and what you’d want to see documented next. I built these labs with the help of AI tools, as part of the broader shift I describe in AI in higher education: What institutions should do next.

Check back here again for more updates.

Frequently asked questions

What is Let’s Do Classwork?

Let’s Do Classwork! is a free set of interactive economics labs created by Mahelet G. Fikru, Professor at Missouri University of Science and Technology. It currently includes seven microeconomics labs and a four-part Environmental Economist Lab, each built from exercises she uses in her own courses.

Are these economics exercises free?

Yes. Every lab is free to use, with no sign-up required.

Where can I practice supply and demand problems online?

The Equilibrium Lab and the Efficiency Lab let you shift supply and demand, solve for equilibrium price and quantity, and calculate consumer surplus, producer surplus, and the effect of a per-unit tax, with instant feedback.

How do you calculate the deadweight loss of a negative externality?

Find the efficient quantity where marginal benefit equals marginal social cost (private cost plus marginal damage). The deadweight loss is the triangle between marginal social cost and demand, from the efficient quantity out to the competitive quantity. The Externalities lab walks through a full worked example.

What is the difference between the free-rider problem and the tragedy of the commons?

Both come from non-excludability. With a public good, people contribute too little, so it is under-provided. With a common-pool resource, people use too much, so it is run down. The Commons Lab works through both.

How much do U.S. manufacturers spend on pollution abatement?

In 2005, U.S. manufacturing plants with 20 or more employees spent about $20.7 billion on pollution abatement operating costs and $5.9 billion on abatement capital expenditures, according to the U.S. Census Bureau’s Pollution Abatement Costs and Expenditures (PACE) survey. Chemical manufacturing had the highest operating costs ($5.2 billion), followed by petroleum and coal products and primary metals. Texas, California, and Louisiana spent the most among states. The Abatement Cost Atlas breaks these figures down by state, sector, activity, and media.

Who are the labs for?

Each lab is labeled by level. Most target graduate applied microeconomics, the Efficiency Lab targets undergraduates, and the Workforce Lab needs no economics background.

Can instructors use these labs in class?

Yes. Instructors can link to any lab from a syllabus or learning management system. Feedback on how you use them is welcome in the comments.

How often are new labs added?

New labs are added as more classroom exercises are converted into interactive versions. Subscribe to get each new lab when it is published.

About the author

Mahelet G. Fikru is a Professor at Missouri University of Science and Technology, where she teaches applied microeconomics courses. Her research covers energy transition economics, critical minerals, and public policy. She writes M G Fikru, a newsletter on energy, natural resources, technology, and sustainability.

In my case, I spent about a year receiving training on online teaching. I completed two formal training programs and earned two certificates. The institution purchased platforms such as Zoom and Panopto. We received guidance on best practices, what to do and what not to do, and even clarification on issues that many faculty were uncertain about, such as whether students could be required to turn on their cameras and what legal considerations applied.

The message was clear: online education represented a major disruption, and the institution had a responsibility to help educators adapt.

My experience with AI has been different.

When AI arrived, everyone responded differently

When generative AI entered the classroom, there was no comparable institutional response. The guidance I received was limited. Faculty were told to use AI detectors, often without institutions purchasing those tools or providing training on their strengths and limitations. We were told to discuss AI at the department level and do what made sense for our disciplines.

In practice, the message often felt like "figure it out."

As a result, I observed a wide range of responses among educators. Some reacted in fear and focused on preventing the use of AI whenever possible. Others largely ignored it and continued teaching as usual, relying heavily on take-home assignments and assessments even as students increasingly submitted those same questions to ChatGPT and received fluent answers. Some faculty embraced AI and began using it to help generate assignments, quizzes, and exams. Many institutions adopted traffic-light policies—green for unrestricted AI use, yellow for limited use, and red for no AI use.

While these approaches provide a starting point, they do not amount to a coherent institutional strategy.

Higher education needs a more structured approach while still allowing flexibility at the departmental and individual faculty levels.

Step 1: Understand AI’s impact on student learning

The first step is understanding what AI is actually doing to learning within our institutions.

Universities should systematically document AI’s impact by gathering information from both faculty and students. Educators should be asked how AI has affected their teaching, challenged their assessments, and influenced their pedagogical decisions. Students should be asked how AI affects their learning, what concerns they have, how they use these tools, and what they expect from their instructors.

Before developing solutions, institutions need a clear understanding of the problems they are trying to solve.

For example, are students completing assignments without understanding the underlying concepts? Are traditional assessments still measuring what educators believe they are measuring? Which learning objectives are most vulnerable to disruption? Where is AI creating opportunities for deeper learning, and where is it creating risks?

Without answering these questions, institutions are responding to AI based largely on assumptions.

Step 2: Move from awareness to action

Understanding the problem is only the first step. Once institutions have documented how AI is affecting learning, they need a framework for responding to it—a framework for action. In my view, that response should focus on three areas:

1. Keep learning objectives at the center

2. Invest in educators

3. Build frameworks, not rulebooks

I am not advocating for rigid mandates or one-size-fits-all policies. Different disciplines have different learning goals, and faculty need flexibility to design courses that fit their contexts. However, institutions can provide a common framework that helps guide decision-making.

Keep learning objectives at the center

One possible starting point is learning objectives.

If the purpose of an assessment is to evaluate a particular learning objective, educators should be encouraged to ask whether the current design still allows that objective to be achieved in an environment where AI is readily available. If AI use prevents students from demonstrating the intended learning, then the assessment may need to be redesigned.

The focus should not be on whether AI is present. The focus should be on whether learning is occurring.

This type of framework would not dictate specific assignments or teaching methods. Instead, it would provide a shared principle that departments and faculty can adapt to their own needs.

Invest in educators

Institutions must invest in educators if they expect meaningful change. During the transition to online learning, universities invested resources, training, and time. Faculty were given opportunities to learn new approaches and experiment with new teaching methods. Incentives were aligned with institutional expectations.

A similar investment is needed today.

If universities expect faculty to redesign courses, rethink assessments, and adapt to rapidly evolving technologies, they should provide the training, resources, examples, and support necessary to do so. Faculty should be given opportunities to experiment, learn from one another, and develop approaches that work in their own classrooms. One example from my own teaching is Let’s Do Classwork!, a free set of interactive economics labs I built from my classroom exercises with the help of AI.

Educators should not be left alone to navigate these changes.

Build frameworks, not rulebooks

What is needed is not a detailed rulebook for AI. The technology will continue to evolve, and today’s solutions may not fit tomorrow’s challenges. Instead, institutions need broad frameworks that help educators make decisions when new technologies emerge.

The goal is to ensure that learning remains at the center of education, even as the technologies surrounding it continue to change.

Key takeaways

During the transition to online learning, institutions invested in helping educators adapt. In the AI era, many educators have been left to navigate the changes on their own.

Universities do not need all the answers. AI is evolving too quickly for any institution to develop a perfect response. But institutions can begin by building systems that help educators respond to change. In my view, understanding AI’s impact on learning, investing in educators, and creating adaptable frameworks are important first steps.

The goal is not to manage AI. The goal is to ensure that learning remains at the heart of education.

Related: The Age of Content Scarcity Is Over. Judgment Is What’s Scarce Now.

When generative AI arrived, I got nothing.

Read more

The Problem

By Mahelet G Fikru

I teach principles of microeconomics using a Cengage textbook. Cengage has built-in interactive homework exercises — multiple choice questions, problem sets, activities — all delivered through their digital platform. Students complete them at home, get immediate feedback, and receive a grade. I thought it was a clean, efficient system. My students thought so too.

What I did not anticipate was how quickly they would find a workaround.

Some students took a screenshot of the problem and uploaded it directly to ChatGPT. Others copy-pasted the question text. Either way, the AI returned an answer — and for multiple choice questions, it was almost always the right one. Students submitted that answer, the platform registered it as correct, and they earned full marks. Most of them got A’s on the homework.

A few weeks later, I gave a closed-book in-class test using the same types of questions. The scores were strikingly low.

What stayed with me from that experience was not the test scores. It was what happened before the test. Some students came to me after completing the homework, holding their A, and asked me why that particular answer was correct. They had the right answer and no idea why it was right. A few told me they had tried to work backwards from the AI output — they took the answer and tried to figure out what it meant — but the logic still wasn’t landing.

They were not cheating, at least not in the way I had understood that word. They were not hiding what they had done. They told me openly, matter-of-factly, as though it were an obvious strategy. And honestly, I could not fully disagree with them. The homework was designed to produce a correct answer. They produced a correct answer. The system rewarded them for it.

The problem was that the assessment was not designed to make them learn. It was designed to produce an output, and AI had made producing that output effortless. The A was real. The learning was not.

That was not a pleasant moment for me as an educator. But it was an important one. It forced me to ask a question I had not seriously asked before: What was the homework actually for?

The Solution

By Devika Toprani, Soulful Learning with AI

What Mahelet describes is one of the clearest illustrations I have seen of what I call the missing cognitive layer. Not missing from the student. Missing from the learning design.

Her students were not confused about economics because they were incapable. They were confused because the sequence gave them the answer before their brains had formed any question. And that is a structural problem. Not a moral one.

There is a name for what happens cognitively when an answer arrives before any internal framework exists to receive it. Automation bias. The documented human tendency to defer to an AI output because it arrived first... and because it arrived confidently (Skitka et al., 1999; Endsley, 2016). The student does not evaluate the answer.

They accept it. Then they submit it. Then they earn full marks. And the gap in their understanding stays invisible.

Invisible gaps do not get fixed. That is the part that does not show up in the gradebook.

The Somagraphic Learning™ framework is built around a three-stage sequence: Attempt → Map → Refine (Toprani, 2026). You can see it in the infographic above and explored in detail on my website. It was designed precisely for this moment. Not as a policy response to AI use. As a cognitive intervention that changes what the learner brings into the encounter with AI.

🖊️ Attempt: Write or sketch what you already think about the question. Messy is expected. Wrong is fine. Incompleteness is the point. Research on productive failure shows that effortful initial retrieval, even when incorrect, strengthens subsequent encoding (Kapur, 2016). The attempt is not a warm-up. It is the first layer of learning.

🗺️ Map: Connect two ideas. Force a relationship. Use the sentence: [this] affects [that] because [your reason]. Even one directional relationship, written in the learner’s own words, changes everything that follows. It creates what I call a cognitive artifact. A timestamped, externalized record of the learner’s own reasoning. Before any AI output exists. That artifact becomes the learner’s reference point.

✨ Refine: Now AI enters. But it enters a different room. The learner is no longer open-handed, ready to receive whatever the machine offers. They are holding their own structure. They are comparing. Questioning. The gap they identified in the Map stage becomes the question they take to AI. That is the difference between generating answers and refining understanding.

Mahelet’s students would have arrived at ChatGPT with something that could not be faked: a directional claim of their own. Something like “demand falls when price increases because consumers have a fixed budget and substitute options.” That one sentence. That one, because it would have changed everything about how they read AI’s response. They would have had something to compare it against.

The Map stage produces what no detection tool, no proctoring system, and no AI literacy curriculum currently on the market produces. Verifiable pre-AI evidence of original human reasoning. Not surveilled. Not inferred. Visible. Dateable. Theirs.

This is also why I built Map Before Machine™ as a deployable tool, not a concept to aspire to. It is printable in under a minute. LMS-ready on Canvas, Moodle, and any comparable platform. It requires no app, no subscription, no curriculum redesign. Ten minutes. A pen. Paper. That is the entire barrier to entry. Because if the solution is not that accessible, it will never reach the students who need it most.

What Mahelet named was the real question educators are circling right now without quite landing on it. What was the homework actually for? Was it for producing a correct answer? Or for building a mind that could recognize one?

The Somagraphic Learning™ framework says: the sequence is the pedagogy. When AI enters, is the design decision that everything else depends on.

The A+ grade was real. The learning was not. And there is a structure that changes that…

Have you seen something similar in your classroom or organization? We would love to hear from you in the comments!

Related: The Age of Content Scarcity Is Over. Judgment Is What’s Scarce Now.

About the authors

Mahelet G. Fikru is an economist and educator who has taught a variety of economics courses for nearly fifteen years. She writes about education, technology, and economics on her Substack, M G Fikru.

Devika Toprani is an independent researcher, educator, and EdTech Futurist with 5+ years of multidisciplinary experience across three countries. She is the founder of the Somagraphic Learning™, a USPTO trademarked, IP-protected, human-first visual cognitive framework that reduces automation bias in AI-integrated learning. She writes at Soulful Learning with AI on Substack.

References

Endsley, M. R. (2016). From here to autonomy: Lessons learned from human-automation research. Human Factors, 59(1), 5–27. https://doi.org/10.1177/0018720816681350

Kapur, M. (2016). Examining productive failure, productive success, unproductive failure, and unproductive success in learning. Educational Psychologist, 51(2), 289–299. https://doi.org/10.1080/00461520.2016.1155457

Skitka, L. J., Mosier, K. L., & Burdick, M. (1999). Does automation bias decision-making? International Journal of Human-Computer Studies, 51(5), 991–1006. https://doi.org/10.1006/ijhc.1999.0252

Toprani, D. (2026, May 8). Somagraphic Learning™ Framework: A human-first, AI-supported visual cognitive approach (Version 3). OSF Preprints. https://osf.io/preprints/edarxiv/fnk7z_v4

I used to assign a straightforward homework question in my environmental economics course:

In your opinion, what are the top three environmental problems in the United States? Explain.

I framed it explicitly as an opinion-based question. And yet, in the age of generative AI (GenAI), that wasn’t enough.

Despite being asked for their own views, many students still turned to AI-generated answers, which I have reflected in my November 21, 2025 post. The responses were polished and largely homogenous. But they revealed very little about how students were reasoning or about how they connected the issue to their own experiences.

In other words, asking for “opinion” did not guarantee that thinking would occur.

So, I redesigned the assignment.

Instead of starting with student opinions, I asked them to first engage with GenAI explicitly:

1. If you ask AI, “What are the top three environmental problems in the U.S.?”, what three items does it list? Evaluate the response by highlighting strengths and weaknesses.

2. In your opinion, what are the top three environmental problems in the U.S.? Discuss and justify.

The grading rubric intentionally separated the tasks:

• Summarize what a GenAI tool identifies as the top problems

• Evaluate the AI response

• Identify three other problems of personal concern

• Justify those choices based on your own experience

Why this redesign works better

I wanted students to be able to use GenAI—but not stop there. If I had simply asked them to identify the top three environmental problems, learning would likely have ended with the AI output.

Requiring evaluation forces students to pause and think about what GenAI produces rather than accepting it at face value. Asking them to identify other problems pushes them beyond familiar, broadly defined categories (e.g., climate change) and helps them focus on problems they personally care about.

What I learned

Students had no difficulty generating AI responses. GenAI reliably produced broad, accurate, and widely accepted categories.

The learning happened when students were required to evaluate those responses and then articulate their own concerns. The class moved from convergence to diversity. Instead of repeating the same polished answers, students engaged with more specific, localized, and less visible environmental issues. Their reasoning was grounded in observation: what they see, what has changed, and what affects their own communities.

GenAI can produce a competent answer to almost any question, including opinion-based ones. But learning happens when students are required to make judgments, explain tradeoffs, and defend priorities. This redesign did not ban GenAI. It made GenAI visible, assessable, and insufficient on its own.

My redesign accomplished what the old prompt could not: it shifted GenAI from an output to be submitted to an input to be assessed. It created a clear difference between a polished answer and a reasoned answer.

Two questions for you

I share this as one example of how assignment design can be rethought in the age of GenAI. I am curious how others are approaching this shift:

1. How are you redesigning assignments?

2. Where are you allowing AI use, and how are you structuring tasks so that learning still requires evaluation, judgment, and personal reasoning?

Let me know your thoughts in the comments.

Read more

Read more

In my previous article, I discussed how student responses to assignments have evolved over the last three years, as AI…

Read more

The short answer: Each year, I ask students in my environmental economics course the same first-week question: what are the top t…

Read more