On Capacity, Sustainability, And Attention

Photo by Nick Fewings on Unsplash

Instead of summarizing the latest AI announcements, I need a reset. Truly, we all do. For the past two years, educators have been asked to reevaluate their teaching and assessments in the wake of ChatGPT, adopt or refuse it, develop policies, and become AI literate. Many of us have tacitly accepted some part of this as our jobs. After all, it’s expected for a teacher in higher education or K-12 to pursue professional development to keep up to date with developments within their field. We go to conferences, workshops, or other trainings. Except generative AI isn’t a normal or novel development within our field of study we can attend some conferences or webinars to understand its impact to keep up with it. None of this has been normal and we need to stop acting like it is.

Attention Is What We Desperately Need

Seven years ago, a group of machine learning scientists at Google proposed this odd contraption called a transformer (the T in GPT). They did so in a paper aptly titled Attention Is All You Need. How ironic it is the title of that paper speaks to our current moment in education. I’m fresh out of attention and time. You likely are as well. The so-called ‘magic’ of GenAI—the ability to let users save time by automating tasks, matters little when so many changes occur so rapidly that our brains disengage and create boundaries.

This past month alone we’ve seen the public release of one AI update after another. Notably, none of these are simple upgrades. Many change the way users interact with data in such profound ways that society will be left grappling with the implications for years. Google’s NotebookLM’s podcast feature suddenly got an update that lets a user pause the recording and talk to the two synthetic voices in real-time. This follows Google’s Multimodal Live launch that lets users share their screen, and their camera, and lets Gemini ‘see’ and ‘talk’ to them in real-time. OpenAI released similar features with their Plus plan, along with the long-awaited video generation tool Sora.

Multimodal voice and vision are just the tame stuff. The next wave of AI tools announced by developers is agentic—allowing an AI to control your browser and make decisions using all the multimodal features mentioned. It can solve complex tasks across domains using a variety of tools on its own. No, it isn’t intelligent, but it is close to the SciFi AI we’ve all grown up imagining our future would hold. You can sign up for many of these newly announced AI agents on Google Labs. No one knows what this means for education or broader society. Even the folks at Google admit they don’t know the lasting impact releasing AI agents into the wild will have on our economy, our culture, and our very sense of reality itself.

The thoughtful AI literacy framework the MLA-CCCC Joint Task Force on AI and Writing proposed with their third working paper isn’t going to be able to offer a response to agentic AI. Nor is the framework set forth by the Refusing AI in Writing Studies movement. Adoption v. refusal misses the point. Regardless of the framework there is a fundamental challenge we cannot overcome regarding engaging AI in education—the capacity to keep up with and process what’s happening. We’re not going to see AI adoption or AI refusal in any meaningful sense until we first address the material conditions of our labor in this generative AI era.

We Need Sustainable Support to Engage AI

It should be valid for a faculty member to demand resources for something that is outside their realm of understanding. Administrators requesting faculty take the time to develop course policies around AI and help students become AI literate should take a long look in the mirror and ask if this approach is sustainable for themselves before asking it of their faculty and staff. I’ll give you a hint—it’s not!

Faculty must have the time and support necessary to come to terms with this new technology and that requires us to change how we view professional development in higher education and K-12. We cannot treat generative AI as a one-off problem that can be solved by a workshop, an invited talk, or a course policy discussion. Generative AI in education has to be viewed as a continuum. Faculty need a myriad of support options each semester:

• Course buyouts

• Fellowships

• Learning communities

• Reading groups

• AI Institutes and workshops

• Funding to explore the scholarship of teaching and learning around generative AI

If you’ve worked in faculty development all of these items are likely familiar to you. Centers for teaching and learning, academic innovation units, along with others have been doing some combination of the above for faculty long before AI and will continue to do so long afterward. But it hasn’t been enough. Institutions must decide to fund faculty professional development in radical ways to meet our moment.

Engaging AI in 2025

This will be my last full post of 2024. I’ll do a recap and a very public thank you post for the support readers have given this newsletter before the end of the year, but that concept of sustainable attention is one I’m going to be considering going into 2025. During the break, I’m going to try and find a way to maintain my capacity related to AI in education. When I can do that, I plan on making a series that embraces that concept of sustainability and capacity, structuring posts that focus on making generative AI manageable and information transferable so faculty can choose to refuse or embrace elements of AI.

I’m tentatively calling this the Engaging AI series. Each post will guide readers through 90 minutes a week of reading, exploration, and reflection. You can break this up into 30-minute segments three times a week—30 minutes of reading, 30 minutes of exploring an AI tool/ use case, ending with 30 minutes reflecting on what this may mean for your teaching.

An hour and a half a week isn’t enough to learn every little thing about AI—that’s not the point. Breaking off topics into manageable chunks of time speaks to our own need to preserve our capacity and parse through the sea of AI hype that these companies include with their marketing. 90 minutes a week is enough to keep yourself up-to-date, informed, and engaged. My greater hope is that more people who take this path use what they learn to help advocate for more time and resources for us all.