An Open Letter to My Colleagues: How Will AI Change Our Work?

An open letter to my colleagues: How will AI change our work?

Dear Colleagues,

I have been writing (open) letters to my students about topics such as reading, writing , and fortitude . My hope is to encourage them to vigorously embrace the challenges and opportunities that lie ahead.

With this letter to you, my fellow scholars working at the University of Colorado and beyond, I hope to do the same by addressing the sudden, striking emergence of generative AI (GenAI). I believe these technologies will quickly redefine the primary aspects of our jobs—research, teaching, and administrative duties—compelling us to rethink how we work.

Arthur C. Clarke noted, “Any sufficiently advanced technology is indistinguishable from magic.” My magical introduction to GenAI occurred early last year. I sat in a café experimenting with an early version of ChatGPT to address a sticky bit of writing. Awestruck by its speed and ability, I began to cry. (Call me a sensitive nerd.)

There is a common quip: “Fast. Good. Cheap. Pick two of the three.” These predictive models, known as large language models (LLMs), can break this so-called Iron Triangle, performing astonishingly well, fast, and inexpensively. (I was using a free version at the time, but have since upgraded to a paid version of Chat and Claude, each just $20 a month).

Much like how the printing press and internet transformed how we access information and communicate, GenAI will transform how knowledge is produced and transferred. But the speed of change will be measured in months rather than decades or years. This letter follows the recent release of ChatGPT4o (“o” stands for omni), which makes the LLM that moved me to tears seem pedestrian.


Highlighting how gatekeeping in academia stifles innovation, Max Planck famously said, “Science advances one funeral at a time.” Of course, GenAI won’t help overcome the outdated beliefs of senior scholars. However, it can strip away other barriers, just as technology has in the past. Scholars used to perform t-tests manually and type up papers on typewriters (and submit manuscripts by mail). With the advent of GenAI, we now have the tools to significantly speed up these processes and a host of others.

This technological leap offers exciting new possibilities, particularly in behavioral research. GenAI will rapidly analyze large datasets, suggest further analyses, and produce tables, figures, and results sections (including presenting the correct degrees of freedom). GenAI can quickly sift through vast literature, evaluate hypotheses, and suggest studies. This efficiency also extends to writing and formatting papers, managing citations, and preparing data for publication, making the entire research process smoother and more productive. (I am working on a paper that outlines how to approach collaborating with GenAI for behavioral research. Hit me up if you want to see a draft.)

By automating repetitive tasks and providing high-quality analysis, GenAI frees researchers to focus on more innovative and creative aspects of their work. However, the integration of GenAI into research presents significant challenges. One major issue is the anticipated surge in AI-assisted research submissions, which will strain our already overwhelmed peer-review system. We will need to (finally) innovate our review processes, possibly integrating GenAI itself, to manage the burgeoning volume of papers while maintaining the integrity and quality of scholarly communication. The potential for GenAI to produce high-quality but unoriginal content means we need to rethink our criteria for novelty and contribution to prevent research from becoming further commoditized.


I am struck by the variance in student’s adoption of GenAI, from those at the leading edge to those who only use it to complete assignments. Creating assignments that are entirely AI-proof will become increasingly difficult and counterproductive. Eventually, we’ll need to accept that students will use AI just as they use the internet. Instead, we should focus on developing assignments that emphasize uniquely human skills such as creativity, critical thinking, and emotional intelligence. For those committed to a pure assessment of knowledge, I suggest bringing back pencil and paper tests. (The good news is that you will be able to take a picture of the test and have AI do the grading for you.)


The integration of GenAI into our teaching methodologies demands a deliberate approach. We must explore how to best leverage AI to augment human instruction. The technology will drive the hybridization of human and AI instruction, requiring us to design courses and programs that thoughtfully integrate AI assistants as an integral part of the learning experience.

I am just beginning to experiment with the technology, but I see it revolutionizing the way we teach. For example, for those of us who “flip” the classroom, we can ask GenAI to give lessons prior to class, freeing valuable time and energy for discussion and case studies.

Administrative Tasks

GenAI has immense potential to assist with the more mundane service tasks associated with running a university or contributing to your field of study. I won’t say much about this, but GenAI has the potential to significantly streamline processes, such as compiling citation tables and summarizing teaching evaluations for performance reviews (perhaps even conducting unbiased reviews entirely), freeing up valuable time for more impactful work. Amen to that!

What to do?

These ideas are just a small slice of the potential changes to the marketplace—scholarly and otherwise. I suspect that the integration of GenAI into our academic endeavors will demand a fundamental reimagining of our roles as educators and researchers. I suggest starting now.

  • Educate Yourself about AI: Take the time to learn about generative AI, large language models, and their capabilities. Experiment with tools like ChatGPT to understand firsthand how they can assist with research, writing, and analysis. Check out the custom GPTs.
  • Rethink Your Research Process: Consider how AI can streamline literature reviews, data analysis, hypothesis generation, and paper writing. Look for opportunities to leverage AI to focus on the more creative and innovative aspects of your research. (Again, I have that paper to help you get started.)
  • Adapt Your Teaching Methods: Accept that students will use AI—perhaps better than you—and design assignments that emphasize uniquely human skills like creativity and critical thinking. Explore ways to incorporate AI as a teaching assistant, such as having it deliver pre-class lessons.
  • Advocate for Changes to Academic Systems: Push for innovations in peer review and evaluation criteria to handle the expected increase in AI-assisted research submissions. Argue for investing in AI tools and training to keep your institution at the forefront. Premium subscriptions should be offered by your employer, just as email or journal access is.
  • Collaborate with Others: Connect with colleagues who are also exploring the effects of AI on academia. Share ideas, pool resources, and work together to develop best practices and guidelines for responsible AI use.

I know how challenging the job is. There is never enough time to get things done. However, I believe with the investment of time, GenAI will make the job easier overall—especially by removing a lot of mundane tasks. Start small, but start soon.

Finally, as GenAI evolves beyond predictive models and into Artificial General Intelligence, we will need to ask surprising questions. I will be out of the game by then, but here are a few: Will we still need humans to conduct scientific research? Will we even need universities? How much better will AI be than reviewer C?

What do you think? Write me back and tell me.


Peter McGraw