A Tech Librarian Explains How to Build AI Literacy

AI literacy is now a must-have skill. Here's how to develop it.

The rise of ChatGPT and other AI programs has borne witness to a burgeoning set of ethical and informational concerns about AI. Recently, some scientists and tech industry leaders published an open letter pleading for the halt of AI development until human infrastructure can better absorb it. Many are already working toward laying this groundwork by developing a set of tools and skills known as “AI literacy.”

Doubtlessly, librarians will be on the frontlines of teaching AI literacy, and indeed, they already are. This week, we sat down with Nicole Hennig, a librarian and expert in user experience and emerging technologies, who is teaching an ALA eLearning webinar on “AI Literacy: Using ChatGPT and Artificial Intelligence Tools in Instruction.” Nicole offers librarians a concrete foothold into the AI and AI literacy space and the resources necessary to keep apace with its development.

Let’s start with the foundation. What is AI literacy? How does it click with or differ from information literacy?

One of the best definitions I’ve seen appeared in a 2020 paper by Duri Long and Brian Magerko. They define it as a set of competencies that enables people to

  • critically evaluate AI technologies
  • communicate and collaborate effectively with AI
  • use AI as a tool both at home and in the workplace.

In addition to those three competencies, I would add:

  • the ability to understand the basics of how AI works, including machine learning, neural networks, and large language models
  • the ability to use AI effectively and ethically
  • the ability to make informed decisions about using AI technologies

AI literacy is closely connected to digital literacy and data literacy because AI uses digital data and algorithms. Librarians have been involved for some time now with teaching various kinds of literacy, like data, digital, and media literacies. Now there’s a growing need for AI literacy to be included with those skills, because AI is becoming more integrated into the tools we use every day. It’s also being used across a wide range of industries and professions, so it’s no longer just computer science experts who need this knowledge. Everyone—children, teens, college students, and the general public—can benefit from this knowledge.

More on this topic: What to Read about ChatGPT

You’ve spent more than a decade as a librarian engaged in user experience and e-learning development. How does AI, such as ChatGPT, fit into your career and the career trajectories of other librarians in similar positions?

The recent explosion of news about ChatGPT has piqued a strong interest in me to learn more about it. Since ChatGPT came out to the general public last November, I’ve been aiming to learn as much as I can. There are so many questions! How does it work? What data was it trained on? What is it useful for and not useful for? What are the primary criticisms of it?

In my previous role as head of UX at the MIT Libraries, I became very interested in emerging technologies. I spent a few years teaching online courses and webinars about them for librarians. In 2017, I wrote the book, Keeping Up with Emerging Technologies, covering how to test and implement any new technology, with a focus on user needs, ethics, and inclusion. Because of this background, my curiosity about AI makes sense. How does it fit into our work as an emerging technology? And with my UX hat on, I noticed that creating a simple user interface (chat) was a major factor in the takeoff of ChatGPT. Now that I’m working at the University of Arizona as an e-learning developer, I’m seeing how AI literacy could and should become a standard part of the literacies that we teach.

So for those who specialize in instruction, user experience, and emerging technologies, this is an especially important literacy to stay current with. It’s also an important literacy for all librarians, because we have the potential to use it in many areas of our work. AI expertise is increasingly being listed as a job requirement in many fields. I agree with those who say, “AI will not replace you. A person using AI will.”

We have a strong need to educate ourselves first, and then educate our users. In the May 17 webinar I’m offering, I‘ll cover not only the basics of generative AI, but also give inspiration and tips for offering instruction to our users.

Librarians, among others, have raised a host of criticisms of ChatGPT, from inaccurate information to infringed intellectual material and so on. How does AI literacy address some of these concerns?

It can feel overwhelming at times. There is so much polarization—with various factions criticizing it for different reasons. The experts don’t seem to agree on anything. And of course there’s the usual amount of hype (it will save the world!) and more than the usual amount of fear (it will destroy the world!). AI literacy can help people cut through the hype and find a balance between polarizing extremes. When you have a better understanding of these technologies, you’ll know how to use them effectively and ethically. You’ll learn what’s important to consider when evaluating them. You’ll become familiar with open source AI tools. You’ll learn of organizations that are focused on addressing bias and false information.

I believe it’s a mistake to focus mainly on issues of plagiarism, bias, copyright, and misinformation. Those are very important issues, but there are also many ways that AI is helping democratize and enable creative content generation. And when you develop a broader understanding of AI literacy, you’ll learn how solutions are being invented to address the many ethical issues. Because of the exponential growth of the research, some of the problems we see today will be solved very soon.

As librarians, we have an important role to play. We can educate ourselves and others, and influence future developments of AI technologies.

In your webinar, you plan to give several examples of how to use ChatGPT effectively. Could you give us a preview of one?

Sure. One task ChatGPT can be good at is brainstorming ideas for teaching.

To get a useful response from ChatGPT, it helps to understand the importance of prompting. If you only ask a simple question, you may not be happy with the results and decide to walk away. This is a mistake. It’s better to follow best practices for writing your prompt. Follow these steps:

  1. Give the context. For example, “You are an expert in [fill in the blank for your topic].”
  2. Be very specific and concrete about what you want it to do.
  3. Give clear instructions on what the output should look like.
  4. Be simple and concise in your language.

Here’s an example you can try:

“You are an academic librarian focusing on instruction. Please make an outline of important topics to cover for a classroom lecture on ChatGPT.  This will be part of a series of instruction sessions about information literacy. You have already worked with this class on the basic concepts of information literacy. Create the outline as a bulleted list and follow it with a classroom activity based on one of the topics you’ve discussed about ChatGPT.”

When you get the results, you can then ask it to give you an additional activity or set of activities on the same topic. If you’re not satisfied with any part of the answer, you can ask ChatGPT to revise it in any way you like. To learn more about prompting, I recommend this free online course, Learn Prompting.

The free version of ChatGPT uses an underlying model known as GPT-3.5. If you pay for the Plus version ($20/month), you can use ChatGPT with GPT-4, which will give you even more detailed answers. It’s a more advanced model, trained with more information. In my experience, it gives better answers than GPT-3.5.

What are the first steps a librarian without a technical background can take to develop a basic understanding of AI and its implications?

I would recommend starting with this short five-lesson email course that was published by The New York Times, How to Become an AI Expert. The email series is over, but you can access it on their website. It does a good job of explaining this topic for beginners. (You may need to create a free NYTimes account.) I think a better title would be “How to Get Started with AI,” because it’s not expert level.

Another very useful resource is Nano Tips for Using ChatGPT with Rachel Woods. This “nano” course from LinkedIn Learning has only ten video lessons, each about one minute long. But what you get from it is excellent. Woods aims this series at people in business, but the lessons make sense for librarians as well. You can finish it in about ten minutes!

Also, since there are so many terms that are unfamiliar at first, it’s a good idea to keep a glossary list for yourself as you learn. Here are two glossaries that are good to bookmark as a starting point: The A to Z of Artificial Intelligence by TIME and Artificial Intelligence Glossary by The New York Times.

You may also find it helpful to meet regularly with a few colleagues to share what you’re learning and trade ideas.

What are some good ways to stay current with AI topics?

Since this field is changing so rapidly, it’s a good idea to follow a few experts who can help you stay current. You can subscribe to their email newsletters or follow them on social media. Here are some of my favorite people to follow:

And of course, you can follow me on Twitter or Mastodon. I cover AI developments in my newsletter, Mobile Apps News. And finally, you might like to read my ChatGPT FAQ, an informal document created for my colleagues at the University of Arizona.

Register for Nicole Hennig’s webinar, “AI Literacy: Using ChatGPT and Artificial Intelligence Tools in Instruction,” on May 17.

Sign up for LibTech Insights (LTI) new post notifications and updates.

Interested in contributing to LTI? Send an email to Deb V. at Choice with your topic idea.