Assessing AI’s Importance in Journalism and Teaching — Part 1

New technologies like AI can roil a newsroom, but one journalism educator plunged in to see how to make the best of it in the classroom. Photo: Frederick Kent/Iowa Digital Library via Flickr Creative Commons (CC BY-NC 2.0).

EJ Academy: Assessing AI’s Importance in Journalism and Teaching — Part 1

By Emilia Askari

Part 1 of a two-part series

The emails about artificial intelligence began around June 2023, sprinkling faculty inboxes at the University of Michigan where I work with a confetti of mixed tones.

We were excited! But we were also appropriately skeptical, analytical and protective of our students and our research data. We saw bias. We saw Big Tech making money on us. We were better than this new technology. Also, to be honest, we were a bit scared of ways that this new technology may be better than us.

By the end of July 2023, administrators at my school announced the creation of UM-GPT, the first university-sponsored “custom” version of ChatGPT in the world. This development sparked a rousing chorus of “Go Blues” and a video explaining the new AI tools, followed by a daunting schedule of training webinars.

 

Faculty were encouraged to

‘update’ syllabuses and lesson plans

to use AI ‘responsibly,’ just weeks

before the start of the term.

 

Basically, faculty were encouraged to “update” syllabuses and lesson plans to use AI “responsibly,” just weeks before the start of the fall 2023 term.

Since then, I’ve learned a lot and I’ve realized that I have much, much more to learn that could be of value to both journalists and teachers — and for people who play both of those roles.

What follows is a list of ideas I’m carrying into my planning while I continue to seek and appreciate more ideas from colleagues like you. In Part 1 of this two-part EJ Academy, we will discuss how to prepare to use generative AI and how students might use AI productively.

 

Preparing to use generative AI

Don’t assume that all young people understand what AI is or that they’re already using it. In my experience, many students have experimented with AI and many have not. Here are the intro-to-AI resources I shared at the beginning of our last term.

Pay for the premium versions if you can, to protect your thought process and your sources. I feel lucky to work at a university that decided to pay hefty group premium fees for such tools as ChatGPT, Dall-E and others so all students, faculty and staff can experiment with generative AI more privately.

Using such a tool as UM-GPT has two purposes. The first relates to privacy: The objective is to block the AI tools from retaining and “learning” from the information that people using UM-GPT put in their prompts.

 

The objective is to ensure that

everyone at the university has

equal access to AI tools, without

having to pay a premium on their own.

 

The second purpose relates to equity: The objective is to ensure that everyone at the university has equal access to AI tools, without having to pay a premium on their own. I think this a prudent and fair approach, and I’m guessing that other universities — and perhaps some news organizations — will also cover premium accounts for their people. At universities, this premium-for-all approach complies with federal laws around student privacy, whereas student and faculty use of non-premium versions of AI may not.

Dip into a few training sessions about using AI in various contexts, including journalism and teaching. I’ve sampled more than a few of them. Among my favorites is the Online News Association’s AI in Journalism Initiative. Seek out direct conversations with colleagues to help you process how AI is changing education, journalism and related fields like information science. Harvard’s Berkman Klein Center for Internet & Society has a lot to offer.

Consider the ways AI can assist with communication across languages, especially if you have students who are non-native English speakers. Real-time translation by AI of captions in video conferences is getting a lot better — though it’s not perfect. In addition, as a first-generation American myself, I have vivid recollections of my father’s stories about the tremendous challenge of showing up on a U.S. campus as an international student who was not comfortable speaking and writing English. As a teacher, I have observed how some international students who hesitate to ask questions have used AI to boost their confidence when posing written questions or opinions.

 

How students might use AI productively

Have students brainstorm to co-create norms about how AI will be used in your course, with your guidance. Many universities have published suggested language for course syllabi; here’s what my university is offering.

Here's a draft AI policy floated a few months ago by the Society of Environmental Journalists’ new events editorial director, Lisa Palmer, who at the time was working at George Washington University and Planet Forward. In my university’s fall 2024 term, I’ve asked students to highlight all passages in their news stories and other assignments that were suggested by UM-GPT.

 

There’s a lot of talk about

how generative AI can write.

For now, I think it may be more

useful as a reporting tool.

 

Use AI to suggest new story angles and sources. There’s a lot of talk about how generative AI can write. For now, I think it may be more useful as a reporting tool. My teaching partner, Julie Halpert, and I assigned our students to use UM-GPT to explore potential new angles on story ideas that the students wanted to develop. We suggested prompts that the students could adapt. We asked the students to share a sample story with UM-GPT and prompt the AI tool to provide missing perspectives and good people to interview.

As a sample of how this exercise can lead to interesting insights, I offered this experimentation where I asked UM-GPT to suggest angles on a potential story about eggs. I thought the results were surprisingly helpful, though they needed further reporting and verification. Here is a student-produced podcast episode about eggs, which was developed (with lots of additional hard work and verification) from the AI experiment above.

Develop sample prompts about writing a news story, then assign students to use AI to write drafts of their stories and turn in the AI versions. The next step for those students: Turn in another draft of the same story, demonstrating how they are capable of producing better news stories than AI. Here is a sample assignment that we used last term, including some suggested AI prompts.

If you have a peer editing assignment (which I highly recommend), ask students to use AI as a comparison editor there as well. Here’s our sample for this assignment.

 

Identifying the most-used keywords in

a text, and anticipating what search terms

someone might use to find your story

— these are strengths of AI.

 

Use AI to suggest headlines. Identifying the most-used keywords in a text and anticipating what search terms someone might use to find your story — these are strengths of AI. Of course, humans should edit AI’s suggestions before posting them.

Require students to formally reflect on their experiences with AI. I’m thinking of asking students to completely avoid the use of AI to write any reflection assignment. We want the student’s thoughts in those reflection assignments, not just some “thoughts” that an AI bot thinks might sound good.

Here’s a small sampling of what different students had to say over the past year:

“I was shocked by how hard it was for UM-GPT to follow my instructions.”

“UM-GPT can be a tool, and might be more helpful for someone who had no idea where to start, but to me having already done a ton of interviews, it's not super helpful.”

“This class is my first time using GPT, so I am uncomfortable with the tool and hate it, to be frank. I don’t think you should teach with it going forward, though I understand that’s an unlikely request at some point.”

In an upcoming Part 2 of this EJ Academy column, we will continue examining AI by discussing how the tool is changing the context of journalism and emerging ways to use AI more effectively.

Emilia Askari, who has a PhD in educational technology, teaches environmental journalism at the University of Michigan. She also advises a news nonprofit, Planet Detroit; judges the Oakes Award for Environmental Journalism; and serves on numerous Society of Environmental Journalists committees. She is co-editor of EJ Academy.

---

* From the weekly news magazine SEJournal Online, Vol. 10, No. 2. Content from each new issue of SEJournal Online is available to the public via the SEJournal Online main page. Subscribe to the e-newsletter here. And see past issues of the SEJournal archived here.