Things to Assign for Homework That Aren’t Papers

Yesterday on Bluesky I saw this cast from a teacher who had a student whose final got flagged as being 70% paraphrased from AI-generated text.

What ensues is an interesting discussion about the pros and cons of AI usage in an academic context, which is obviously one of the biggest questions facing educators today.

As of the fall of 2024, seven in 10 teenagers in the United States have used generative AI tools, according to Common Sense Media. But most of their parents don’t know. This so-called “homework apocalypse,” a phase coined by Ethan Mollick back in 2023 continues to rage on.

In this piece, Ethan calls out four possible options for how to address the “essay in the room” question. These included:

1. Back to in-class essays.

2. Keep outside of class essays, and forbid AI use.

3. Keep outside of class essays, and encourage AI use.

4. Embrace flipped classrooms (instruction is done by watching videos/AI tutors/readings outside of class, class is for activities and active learning).

   See the full list, and Ethan's full essay here.

Unsurprisingly, what we're seeing (just like that teacher who posted on Bluesky) is that teachers who forbid AI use for outside-of-class essays are finding it increasingly difficult to maintain this approach. Not only does it risk diminishing student learning outcomes, but it also shifts the focus from evaluating content to policing AI usage. This sounds frustrating and unsustainable for everyone.

A few years ago, when I invited an AI founder to visit a high school with me, and a similar question from teachers came up, he offered the provocative thought:

While this obviously invites a host of other problem (such as: how would a teacher have the time to read 30+ full books to evaluate these projects), it does make a compelling point: We aren’t thinking big enough.

Instead of using AI to enhance the small, familiar tasks we’re accustomed to, why not embrace its potential to tackle projects that wouldn't be possible otherwise? These so-called AI-Supercharged Projects could challenge students to think bigger and more creatively.

While traditional essays already assess skills like constructing compelling arguments and engaging an audience, incorporating AI tools into assignments allows students to explore new dimensions of creativity, media literacy, and interdisciplinary thinking. By evaluating how effectively students use AI to enhance their ideas, package content into diverse formats, and present their vision from multiple angles, we can foster innovation and critical thinking alongside foundational skills.

After all, projects are just homework for grown-ups. That AI tools lower the barrier to entry for anyone means that we can invite students to create micro-projects and holistic experiments at earlier ages.

For example, let’s imagine I’m teaching a class that involves assigning the dystopian book, 1984 by George Orwell. Here are some ideas for some AI-Supercharged Projects as homework that you might consider assigning as an alternative to a final essay or paper.

1. Create a fan fiction screenplay, set in the modern era. Lean into the generative nature of AI tools and encourage students to take their favorite scene or vignette from the book and use an LLM (like ChatGPT or Claude) to come up with an entire screenplay. Students would be evaluated on their ability to effectively prompt the AI with relevant details and information that most closely resembles how these characters might actually act. This project would combine literary analysis with creative storytelling skills.

2. Make a comic book training manual for the “Thought Police.” For students with more of a creative bent, encourage them to design a step-by-step training manual for new recruits of the Thought Police, using an AI-powered image generation tool, such as AI comic factory. Students could be evaluated on their ability to pick up on key psychological and ethical components of this role, and also on their close reading skills of examples of how the Thought Police are taught to maintain control in the book.

3. Produce an entire podcast season about the key themes from the book. Drawing upon the literary themes explored in class, instruct students how to collate their own notes with critical literary analysis in other sources to create a series of NotebookLM projects. Convert each into a podcast and then produce the podcast as an entire season that explores various elements of the book. This project would teach students how to curate information effectively and extract the most relevant and universal themes.

4. Create an AI-powered mock propaganda as “Big Brother.” By leaning into AI-powered video creation tools like RunwayML, image generation tools like DALL-E, Midjourney, and Flux, encourage students to design their own end-to-end “mock propaganda” campaign, leaning into themes from the book. Evaluate students with an in-class debate on what they learned about how easy it is to create faux-media, and recommendations for how they can fine-tune their own radar against false information.

5. Create a dystopian society simulation. Using AI-driven platforms, students can simulate a society inspired by 1984, exploring the dynamics of totalitarian regimes and the impact on individual freedoms. This experiential learning activity fosters empathy and a nuanced understanding of the novel's themes.

6. Design an AI-Generated art exhibit. Students can use AI art tools (or even web3 native creator tools and generative art platforms) to create visual representations of the novel's themes, characters, or settings, culminating in a virtual or physical art exhibit. This project integrates literary analysis with artistic expression.

7. Develop a Newspeak language app. Challenge students to create a simple AI-powered app that translates text into Newspeak, exploring the novel's linguistic themes and the concept of language as a tool of control. This project combines coding skills with literary analysis.

8. Organize a virtual debate on surveillance. Facilitate a virtual debate where students use AI to research and present arguments on contemporary surveillance issues, drawing parallels to the novel's depiction of Big Brother. This activity enhances critical thinking and public speaking skills.

By the way, those first four ideas came from me. The second four came from ChatGPT. (Just in case you’re keeping score.)

I know you're probably thinking that these sound great but would never work, really, in a practical setting or a real classroom. You're probably right. In an AI-native era, ideas are cheap; execution is all that matters.

There are many issues to unpack—issues like rethinking evaluation, outcomes, and ethics. We also need to make sure we're not asking too much of already overburdened teachers by expecting intricate experiential learning for every assignment. These solutions need to be sustainable. Practical. Pedagogically sound. Affordable. Accessible. Oh yeah, and students still need to learn how to read. (And I know this alone doesn't come without its challenges.)

But maybe there's something to this idea of blowing up homework. After all, if we want the next generation to think bigger than we do today, that mindset needs to start in the classroom. With AI, no problem is too small. And no idea is too big.

So I wonder... what would an AI-supercharged assignment look like in your field?