Notes from Launching an AI Journalism College Course in 2024
Nov. 26, 2024
Editor’s note: Like many of you, I’m spending more of my social media time on Bluesky at Mikereilley.bsky.social. Follow me there for more tips and tricks. I’ll remain on X/Twitter (for now) but will be sharing more resources on Bluesky. We’ll be adding Bluesky resources to our social media page in the coming days.
I am a week away from completing my first semester of teaching an AI Journalism undergraduate course I created at the University of Illinois Chicago, where I’ve taught digital and data journalism courses in the Communication Department for the past decade.
I spent several months designing the 200-level course (18 students), one of the first of its kind in the country. Like most of my classes, the students publish work to The Red Line Project, a website that covers news and urban issues in Chicago neighborhoods. We built a dedicated AI Storytelling page that features midterm and final project work.
The course will be back for the spring semester of 2025 and hopefully beyond. Here’s what I’ve learned from planning and executing the course:
Book the class loosely — especially on the back end — as technology evolves during the semester: University of Nebraska-Lincoln journalism professor Matt Waite taught an AI class last spring and was one of the first people I called when developing my course. His advice: Book the class loosely as the technology changes rapidly during the semester. He was right. I booked the second half of the semester loosely to adjust to new tools as they became available. I told the students this from the start and it worked well for them.
Teach concepts, not tools: The students love learning the AI software for practical application, but I frontloaded the class with theory, law, ethics and fundamental concepts. We also talked about the history of AI, how to structure prompts and how to attribute its use in a story. We examined biases, how to detect deepfakes and potential threats — and benefits — to society.
Get the students involved: I had each student present an AI tool once throughout the semester. The tools ranged from homework help, reporting, editing an research tools, to how Spotify has implemented AI into its virtual DJ. It’s fun watching the students learn from each other. Pro-tip: Build a working list of links to what they presented and give them the archive at the end of the semester.
Set up clear guardrails and best practices, especially early: We discussed The Red Line Project’s AI policy and how it applies to our classwork in week one. We brought in Lynn Walsh of Trusting News to discuss how to disclose AI use in your story.
Don’t assume the students know more than they do: Outside of using a few simple chatbots, my students had little knowledge how to use LLMs, write prompts and test limitations of the tools. We started there in week 2: how to structure a prompt and tailor it to fit your needs.
I showed them how to use it for research but not as a traditional search tool. We started with Google’s GenAI search and how it has evolved from link-driven search to answer-driven search. Then we moved to Perplexity, custom GPTs and other tools.
Be fearless, but be smart about it: On day one I told the students we would experiment with tools in here and fail with some of them. We did just that when testing image generators, analyzing data, building video and writing prompts. The key: what did we learn from our mistakes? It’s not a failure if you learned something from it.
Experiment! I reached out to some AI companies and offered them feedback from the students and I in exchange for using their tool. My students used Nota to create social media videos off their stories. They also used Rolli to find expert sources and spot troll accounts on social media during the election. (Full disclosure: Both advertise on my site and this newsletter)
They also used Google’s Notebook LM to create podcasts and FAQs from their midterm essays. ChatPDF became a popular tool for quizzing course materials before exams after one of the students shared it in class.
We stressed AI editing but not writing outside of basic grammar, style, punctuation, etc. detection. We used Eleven Labs to record our writing and play it back to us, then make editing changes based on what we heard as well as read. We also built unique audio voices in Eleven Labs and had them read local news stories.
Looking ahead: This spring, I want to develop a concept and build a custom GPT and/or chatbot that benefits journalists. I’m also sitting on a UIC communication department faculty committee that’s developing best practices for professors working with students and AI. I’ll share more about what I learn in a late spring newsletter.
If you are teaching and need ideas, just drop me a note. I’ll share my syllabus and ideas with you, but not my course materials.
Sponsor
Try out Rolli’s new Slack integration!
Google’s Learn About Tool
I’ve just started tinkering with Google’s new Learn About, a conversational learning companion that “adapts to your unique curiosity and learning goals.” I’ve found the tool fun to use for day-to-day tasks such as “tell me about the history of writing” and the like.
The tool is intended to help us grasp new topics and deepen our understanding by packaging material together in a logical way. It goes beyond traditional link-driven search and provides context, but presents it in a more digestible way than LLMs such as Gemini or ChatGPT. In the example above, it offers pictograms and ideograms and other suggested topics in the upper left.
I like its simplicity and will continue to share ways we can use it journalistically.
Full disclosure: I’m the lead trainer for the RTDNA/Google Fact-Checking Training Program.
Sponsor
Be sure to check out the incredible production tools suite with our new sponsor at HeyNota.com
New in AI tools:
MIT AI Risk Repository
Detailed list of 700-plus risksFLUX Image Generator
Free FLUX text and image generationPhotoroom
Mobile and desktop AI tool that creates imagesMeco
Read and discover newsletters
Become a sponsor: Would you like to sponsor future Journalist’s Toolbox™ newsletters? Our rates are reasonable and our sponsors get results. Contact me at mikereilley1 [at] gmail [dot] com
ONA AI in Journalism Trainings Extend Through the Fall
You can still sign up for the Online News Association/Microsoft AI in Journalism Initiative training program, which is continuing through fall 2024 with more sessions. The trainings are free for ONA members and $25 for non-members. You can register for the trainings here.
The trainings, which last 60 to 90 minutes, will run each month with a few extra trainings and mini-labs sprinkled in. They will be recorded for those who miss the live Zoom trainings or want to replay them. I helped them design the trainings and will conduct several of them.
Textbooks
The Journalist’s Toolbox
My new book, “The Journalist’s Toolbox A Guide to Digital Reporting and AI” was published by Routledge in December. You can order it here,
Data + Journalism
Samantha Sunne and I co-authored a textbook, “Data + Journalism” that’s available now on Routledge. (Order here). It’s an introductory- to intermediate-level guide to learning data storytelling from A to Z.
In Quotes …
"Technology alone is not enough. It’s technology married with the liberal arts, married with the humanities, that yields us the results that make our hearts sing." — Steve Jobs
Follow me @itsmikereilley | @journtoolbox | Subscribe on YouTube | Subscribe to this newsletter
Copyright and Trademark 2024 | Mike Reilley