The November 2022 arrival of ChatGPT – an interactive artificial intelligence writing tool that generates natural language text-based responses to user questions and prompts in seconds – made an largely invisible technology (artificial intelligence) highly visible to millions of users.
Artificial intelligence (AI) tools are already prevalent in our lives - including recommendation algorithms on social media and streaming platforms, facial recognition used to unlock phones to autocorrect, and autocomplete features in email and text messaging apps. But AI tools often remain invisible - meaning that we don't spend a lot of time thinking about what they are and how they work (maybe we should!).
ChatGPT (meaning "generative pre-trained transformer”) meant people began taking notice of AI tools. When launched publicly, it gained more than 1 million registered users in just 5 days - becoming the fastest-growing technology platform ever (Harris, 2022).
ChatGPT functions as a digital writing producer for users. Almost instantaneously, it circles through its database of text to produce natural sounding sentences and paragraphs in response to a prompt from a user. It can write essays, draft emails, create a list of ideas for a birthday party, write lesson plans, offer step-by-step tutorials for solving math problems, write blog posts and articles, draft a cover letter or resume, suggest life hacks or financial tips, describe complex scientific processes to different age levels (e.g., describe string theory to a 5 year old), write code, and create content in multiple languages. It can even write a poem about golf in the style of Maya Angelou!
Such results are amazing, blurring the lines between what humans can do that machines cannot.
Take this quiz from the New York Times to see if you can tell if a fourth grader or the new chatbot wrote an essay.
And, imagine what might happen in a future where news articles and media materials are written by AI...wait you don't have to imagine that, it's already happening (see CNET found errors in more than half of its AI-written stories).
But, what are the implications for everyday citizens when politicians, campaign managers, and political staff can have AI writing tools draft press releases, campaign slogans, social media posts, blogs, and more?
- Will you, or any other citizen, be able to tell if what you are reading from a political group was produced by AI tools and not actual human writers?
- What might happen when voters find out that everything written by a politician they were backing was generated by AI?
And, what impact might ChatGPT have on social studies/history, civics, and government learning if students turn to this tool (known for spewing misinformation) for help with their research?
ChatGPT brings with it real complexities and dangers. Technology researcher and educator Torrey Trust has shown that the system is not always safe or reliable as a information source (see ChatGPT & Education slide deck).
If AI tools produce false or misleading claims, then people’s roles as democratic decision-makers are curtailed, distorted, even blocked. In this context, how do we learn to analyze what we are reading to distinguish authentic from false, and AI from human?
Activity 1: Critically Interrogate ChatGPT Text
- First, complete the "Design an Online Political Campaign for a State Political Office" activity on your own (or in small groups).
- Review your classmates' political campaigns and vote on the most influential one.
- Then, as a class, ask ChatGPT to generate all the text for an online political campaign that matches the one voted as most influential.
- For example, you might ask ChatGPT to generate a series of social media posts related to the campaign promises, write a script for a podcast or video that will be embedded into the campaign website, write all the text for the campaign website based on specific political views and goals, and even devise a strategy for increasing voter engagement with the campaign.
- Next, use the Teacher and Student Guide to Analyzing AI Writing Tools to critically interrogate the text that ChatGPT produced and compare and contrast that text with the text produced by the student(s) who designed the most influential campaign.
- Based on your analysis, design a tool (e.g., flowchart, blog post, rubric) to help your family members tell the difference between AI-generated text and human-written text.
Activity 2: Explore Black History with ChatGPT
The "Exploring Black History with ChatGPT" choice board offers educators and students a way to develop their critical media literacy skills as they investigate text from ChatGPT and explore Black history in the United States (click here to make your own copy of the choice board to remix/use). Given ChatGPT’s potential for generating misinformation and biased information, these activities offer students a critical way to investigate how ChatGPT writes about Black lives and history, and potentially consider why this may be.