AI, ChatGPT, and Other Generative Tools: For Faculty: Teaching and Managing AI Use

How Teachers Use AI in the Classroom

These suggested assignments are from OpenAI and are examples of how teachers have used ChatGPT in their classroom:

Role Playing Challenging Conversations

Students use ChatGPT as a particular persona - such as a debate partner that will point out their weaknesses in an argument, a recruiter that is interviewing them for a job, or a boss that provides them feedback in a particular way. This is intended to help students understand their material in a new way.

Building Quizzes, Tests, and Lesson Plans from Curriculum Materials

ChatGPT can be used to craft quizzes, tests, and lesson plans. Share curriculum with ChatGPT and ask for a fresh quiz or lesson plan that uses modern or relevant examples. Take the ideas and make them your own.

Teaching Students About Critical Thinking

Remind students that AI is not always accurate and to confirm information through primary resources. Assign a task that uses ChatGPT or other AI that requires students to check information for accuracy. Doing this can help students work on their critical thinking and problem solving.


These assignment suggestions are from the Connecticut College:

ChatGPT Query Logs

Ask students to "play" with an AI bot such as ChatGPT or Bing AI by providing it with different prompts.  Students can keep a "query log" of all of their prompts and can reflect on AI's responses in this ongoing journal.  You can provide specific reflection and critical thinking prompts or leave it open-ended.

Revising AI Output

Have students revise AI output from any generative or AI writing app (ChatGPT, Bing AI, Paragraph AI, etc.).  Students can work to fact check these tools and provide citations where necessary or they can revise the style choices of the writing.

A List of Possible AI Assignments:

  • Ask students to identify statements of bias generated by the AI tool and have a discussion surrounding the potential source of this bias and its impact.
  • Challenge students to find a question that ChatGPT doesn't want to answer.  Morally charged questions, for instance, often stump generative AI tools.  Have students continue to revise the prompt until the AI answers the question and then reflect on the eventual answer.
  • Ask students to generate policy surrounding AI for different sectors of industry (ex: higher education, business, non-profit work, etc.)

 

MLA and APA AI Citations

Students should potentially be citing any AI tools that they are paraphrasing, quoting, or incorporating into their work. You can share these tips with your students if AI generated work is allowed in your assignments.


MLA Template

Author: We do not recommend treating the AI tool as an author. This recommendation follows the policies developed by various publishers, including the MLA’s journal PMLA

Title of Source: Describe what was generated by the AI tool. This may involve including information about the prompt in the Title of Source element if you have not done so in the text. 

Title of Container: Use the Title of Container element to name the AI tool (e.g., ChatGPT).

Version: Name the version of the AI tool as specifically as possible. For example, the examples in this post were developed using ChatGPT 3.5, which assigns a specific date to the version, so the Version element shows this version date.

Publisher: Name the company that made the tool.

Date: Give the date the content was generated.

Location: Give the general URL for the tool.

Taken from The MLA Style Center


APA Template

Author: The author of the model is OpenAI.

Date: The date is the year of the version you used. Following the template in Section 10.10, you need to include only the year, not the exact date. The version number provides the specific date information a reader might need.

Title: The name of the model is “ChatGPT,” so that serves as the title and is italicized in your reference, as shown in the template. Although OpenAI labels unique iterations (i.e., ChatGPT-3, ChatGPT-4), they are using “ChatGPT” as the general name of the model, with updates identified with version numbers.

Source: When the publisher name and the author name are the same, do not repeat the publisher name in the source element of the reference, and move directly to the URL. This is the case for ChatGPT. The URL for ChatGPT is https://chat.openai.com/chat. For other models or products for which you may create a reference, use the URL that links as directly as possible to the source (i.e., the page where you can access the model, not the publisher’s homepage).

How to Cite ChatGPT, by McAdoo, T. (2023, April 7).  

Assignments to Reduce AI Use

If you prefer not to use AI in your classroom, it may be best to leverage and employ other types of assignments that discourage the use of AI.

Assign Projects

Assigning projects that are completed in steps can help students to use their knowledge in real-world, hands-on settings. It requires students to apply knowledge in a practical matter and AI has a harder time providing those answers. Encouraging critical thinking and problem solving through these projects can ensure that students are grasping concepts.

Open Ended Questions

Using open ended questions requires students to use critical thinking and they cannot rely on AI-generated responses. They can be tailored to student's experiences, interests, and background, requiring to rely on their own knowledge. Asking them to correlate concepts to their personal experiences makes it so they cannot rely on an answer from ChatGPT or other AI source. Personalizing assignments makes it a challenge for students to use pre-generated content.

Students Upload Their Work

You can assign file upload questions from other sources. Upload a picture or document that contains elements from work they have done towards their exam. 

Use Timed Assignments

Time limits can keep students from using AI tools. Limiting their time to respond to questions makes it challenging for students to rely on AI in order to reach an answer. It helps them to focus on understanding and applying materials rather than seeking out a shortcut.

Frequent Assessments

Regularly assessing students can reduce the temptation to cheat. It students are submitting work throughout a course, it can sway them from procrastinating and prepare them for the big exam. 

Follow Up on Assignments

  • Have students read a line or two from the essay and ask them to explain what they mean. 
  • Read separate points from the essay and ask them to explain how they are related.
  • Ask students to summarize points of the essay in a single sentence.

Conduct Research

Have students work on research assignments that requires socialization - interviews, surveys, and fieldwork. This requires knowledge and inquiry, rather than utilizing AI tools.

Verifying AI Reponses

When students are using AI, they may need to utilize resources to check the credibility of AI responses. These are suggested ways you can ask students to do this:

  • Lateral Reading - Lateral reading is the method of looking to see if other reliable sources can be found that valid the information you are unsure about. This could be as simple as a Google search to verify that someone mentioned by the AI bot does in fact exist, or finding validating information on a Wikipedia entry or news article. When the information is confirmed through multiple sources then you can be more confident in the information you are seeking to utilize.
  • Ask the AI for links or sources - If the AI bot is telling you something that you want to verify, you can ask it to provide a source or link and then check that link or source, because current AI models can make up sources. If an article, report, or other source is mentioned then you can seek out that resource (in Google, Google Scholar, or a Library Database) to ensure that it does exist and then to check the content to confirm that it matches what was shared by AI.

Use MSJC Resources to Verify Responses

MSJC Library also offers resources that can be used and assigned to students in order to verify the responses they have received from ChatGPT and other Generative AI. Encourage students to use these sources if they plan on using AI so that they can validate/invalidate information that has been fed by ChatGPT and other Generative AI. This can help students to understand how AI can be used rather than solely depending on the AI's complete response.

  1. Credo Reference: Contains content from hundreds of reference books covering every major subject with detailed entries and images (art, diagrams, interactive maps and photos).
  2. Gale eBooks: Collection of academic eBooks from a variety of subject areas, including encyclopedias and other specialized reference sources.
  3. Oxford Reference: Contains specialized dictionaries and subject encyclopedias written by experts and published by Oxford University Press.
  4. American Indian Experience: Gives exclusive focus to the historically underrepresented realities of indigenous peoples in the United States. Provides both primary sources and secondary sources. Topic-oriented scholarly essays address sensitive topics such as cultural appropriation, etc.
  5. Asian American Experience: Explore the social, political, and economic forces that have shaped the experiences of Asian Americans in distinct historical eras. Includes primary sources, including images, audio clips, documents, and video.
  6. Latino American Experience: Explores the rich heritage and current culture of Mexicans, Puerto Ricans, Guatemalans, Cubans, Dominicans, Colombians, Ecuadorians, and other Hispanic cultures in the United States.
  7. Music Online: Garland Encyclopedia of World Music: Comprehensive encyclopedia devoted to music research of all the world's peoples. Divided into geographic regions. Includes audio files with regional music recordings.

ChatGPT Detectors & Tips for Detecting

ChatGPT detectors claim reliability, but there is no foolproof tool that can reliably tell you how content was generated. They should not be relied on 100% of the time. Turnitin has led to false cheating accusations as it incorrectly flagged students' work as AI-generated and does not create mutual trust between students and teachers.

Educate and stress on your students the importance of understanding how AI is not a reliable tool. It could be addressing AI in the classroom during the first week or two of class. Have them use the tool to understand how it can negatively impact their experience in your classroom. The more they are exposed to its misuse, the better scholars students will become. Additionally, if students do use AI to write assignments, they need to understand the implications of not citing ChatGPT as a source. Encourage students to cite ALL sources, including AI tools. 

Although, ChatGPT detectors are not 100% perfect and should not be relied upon, it can be used to intiate a conversation with students that you suspect could have cheated. Always be sure to confirm assignment details with the student. You could have students utilize these ChatGPT detection devices to understand how their writing is being analyzed. Consider these programs as tools and compare a students work with various detectors, but do not lean on them heavily. It is better to build trust between students and faculty. Check out the AI detector options and how they compare to one another:

1. Originality.AI

Originality.AI is a paid tool described as “the most accurate AI content detector and plagiarism checker built for serious content publishers.” It costs $.01 per credit, which scans 100 words. It also features a plagiarism scanner alongside the AI scanner.

2. Writer

Writer offers a free detector that works on up to 1,500 characters and requires API access for more. This tool essentially thought all of the submissions were relatively likely to be human-generated.  Writer is an AI writing platform, and the free content detector also allows you to input a URL to check.  There is no plagiarism feature within this tool directly, and it gives you an overall score but doesn’t mark specific sections as likely or unlikely to be AI-generated.

3. Copyleaks AI Content Detector

The Copyleaks AI Content Detector has a few interesting features:

  • It gives you an overall verdict (human or AI).
  • If you hover over specific text, it will give you a percentage likelihood that the text is human versus AI copy.
  • It pre-loads with examples across different GPT and human outputs to show you how the tool works
  • There’s a Chrome extension.

The tool is free, and while a single scan doesn't do both, there's also a plagiarism detector.

It accurately identified much of the human-generated content as human and much of the AI-generated content as AI. However, it wasn't always confident and was incorrect in some individual sections.

4. OpenAI's AI Text Classifier

AI Text Classifier is a free tool from OpenAI, the company that created GPT (which is what runs many generative AI tools) and ChatGPT. If you thought the company that makes the most popular AI tools would have the most accurate detector, that didn't ring true. The tool identified the human content as “very unlikely” to be AI-generated but identified the content from OpenAI’s own platform as “unclear” if it was AI-generated and “unlikely” AI-generated with and without the additional prompt. The tool is free (you need an OpenAI account). There aren't a lot of bells and whistles and no plagiarism check. You get a finding but no precise score. It also won’t highlight specific sections of copy for their individual likelihood of being AI-generated or not.

5. Crossplag AI Content Detector

Crossplag offers a few free scans without an account. When logged in, you'll get unlimited free scans. The tool does give an overall score but doesn't offer the ability to check for plagiarism or mark up individual sections of the content. Crossplaq identified the human and AI content as 99% human. It is worth noting that the tool repeatedly calls out that it performs better on longer text.

6. GPTZero

GPTZero has free and paid versions. The free version gives you a score and has some pre-loaded examples.  The paid version also offers a plagiarism checker, highlighting AI content, and the ability to bulk upload multiple pieces of content (which is actually designed for a classroom environment, like many of these tools). Pricing for the paid tools isn't listed on the website, but you can fill out forms to request access to the paid tool and the API.

7. Sapling AI Detector

Sapling AI Detector offers free and paid versions, a Chrome Extension, an API, an overall score and sentence highlighting. Sapling identified the human content as mostly human and the AI content as almost entirely AI. However, it identified the “perplex” and “bursty” ChatGPT content as highly likely to be human.

8. Content at Scale's AI Detector

The Content at Scale AI Detector is free and offers a score with a breakdown of different elements within the content (predictability, probability, and pattern) and pre-populated examples. No plagiarism detection option exists, and individual content sections aren’t highlighted.

Content at Scale rated all three samples as 92% human or higher. The human did have the highest human percentage, followed by the prompt-modified GPT content and the content generated by the simplest prompt.

9. ZeroGPT

ZeroGPT is a free tool that offers an overall score and highlighting but no plagiarism detection. The tool returned low percentages of AI-generated likelihood across the board. It roughly corresponded to the human-ness of the samples, with the human-generated sample getting the lowest AI percentage, followed by the burstiness / perplexity prompt, and then by the simpler prompt).

10. GLTR

GLTR, or the “Giant Language model Test Room,” is a free tool that was a collaborative effort between Hendrik Strobelt, Sebastian Gehrmann, and Alexander Rush from the MIT-IBM Watson AI lab and Harvard NLP. The tool offers some overall scoring but mainly shows whether each word within a text was one of the 10, 100, 1,000 or outside the 1,000 most likely words to be generated by AI based on the prior context in the text. There is no plagiarism feature set here. The tool returned the highest number of “likely words” and particularly likely words in the top 10 for the text generated by the simpler prompt. Still, the distribution was similar for the human and additional prompt copy in this small test.

11. ChatGPT Detector on Hugging Face

This is a free detector that gives an overall prediction and score. It doesn’t have any plagiarism or highlighting features. This tool generated over 99.9% likelihood of human content for all three samples. 

12. Corrector AI Content Detector

Corrector AI Detector is a free tool with a 600-word count limit and a percentage score. There is no plagiarism or highlighting features for the tool.

The tool labeled each of the samples with a very low percentage probability of being “fake” or AI.

13. Writefull GPT Detector

The Writefull GPT Detector is a free tool with a simple score and API access. There is no highlighting and no plagiarism detection. It found all three samples as likely human but did have the ChatGPT content with the simplest prompt with the highest AI likelihood (19%).

14. Hive Moderation’s AI-Generated Content Detection

The Hive Moderation AI-Generated Content Detection tool is free (with character limits and requiring sign-in after several uses). It doesn't offer plagiarism or highlighting but does allow you to break content out by sections and get segment-specific scores. Hive essentially scored each writing sample properly, with a 0% likelihood of being AI for the human content and 99.9% scores for the AI samples. 

15. Paraphrasing Tool AI Content Detector

The paraphrasing tool is a re-writer, and they offer this free AI content detector with an overall conclusion and highlighting and no plagiarism features. The Paraphrasing Tool concluded that each of the text was likely human.

16. AI Writing Check

AI Writing Check is a free tool that offers a score, has a word maximum, and does not feature text highlighting for AI probability or plagiarism features. AI Writing Check identified all three writing samples as human.

Key learnings from testing 16 AI and ChatGPT content checkers

Again, it’s essential to caveat key findings here because three short writing samples are very small samples to draw steadfast conclusions about the individual tools.

That said, there are a few interesting patterns related to AI writing detectors in general:

  • Calibration: Most tools found all three writing samples either highly likely or highly unlikely to be human. Different tools will likely be “harder” or “softer” graders when it comes to scoring content human versus AI, so understanding how a tool is calibrated can help determine how useful they are.
  • Rarest features: A few rare features across the tools were things like bulk uploads, plagiarism detection, having a Chrome extension, and highlighting specific sections with likelihood of being AI versus human. 
  • Free vs. paid: All of the tools featured here but Originality.AI had at least free versions. But the tools with the most accurate results and most “rare features” tended to be the tools with a paid version.