This resource page was developed to help faculty learn about Generative AI and its
implications for teaching and learning. The information and recommendations on this
page will be updated as more is learned about this technology.
ChatGPT and GenAI FAQ
ChatGPT is a Generative Artificial Intelligence (AI) chatbot that uses machine learning
models to produce human-like text in response to a question, prompt, or instructions.
It was trained on large amounts of text data and can generate natural language text
in response to user prompts. ChatGPT is optimized for generating responses in the
form of conversations.
ChatGPT (GPT 3.5) was trained to write using content that already exists on the Internet
up to the year 2021. Whereas ChatGPT 4.0 was trained on a larger and more diverse dataset, including
additional sources like images, audio, and code from real-world applications. It generates
responses to questions and prompts by predicting the next words or phrases that are
most likely to follow based on the context of the conversation. Users type a question
or prompt in the text box and then ChatGPT produces a response. Users can continue
the conversation by typing follow-up questions or prompts. ChatGPT remembers what
was previously discussed and produces more text. It recognizes, summarizes, and predicts
text based on its giant data sets.
Go to the ChatGPT website, create an account, and start typing your prompts.
If you ask a question, ChatGPT will answer it. This is like giving a command to Google.
Google produces sources for you to extract the information yourself while ChatGPT
produces a written-out answer. Unlike a Google search, ChatGPT can conversationally
generate full paragraphs of text.
ChatGPT can easily write content based on a prompt. For example, you can ask it to
generate a piece of text, like writing a poem. The prompt can include the length and
writing style of the piece. Also, you can ask for sources to be cited and even specify
that you want only ‘real sources’; however, be aware that these will not always be
accurate.
You can provide ChatGPT with pieces of text and ask the chatbot to summarize, paraphrase,
translate, remove spelling mistakes, give feedback, and so on.
If asked to perform multiple tasks at once, ChatGPT struggles to prioritize them,
which might lead to a decrease in effectiveness and accuracy. It performs best when
given a single task or objective.
ChatGPT might produce references that do not correspond to the text created. It might
give fake citations that mix real publications from multiple sources.
The free version of ChatGPT was trained using data from up to the year 2021. This
means that the chatbot (GPT 3.5) cannot interpret more recent data.
ChatGPT might sometimes generate nonsensical or inaccurate responses to certain questions
or situations. It cannot detect subtle emotional cues or respond appropriately to
complex emotional situations.
Responses may not be entirely accurate in terms of context or relevance, providing
incorrect, oversimplified, unsophisticated, biased, inappropriate, or irrelevant content.
The trained data may contain biases or prejudges. ChatGPT may sometimes generate responses
that are unintentionally biased or discriminatory.
ChatGPT cannot access all the knowledge that humans possess. It may not be able to
answer questions about very specific or niche topics. Also, it may not be aware of
recent developments or changes in certain fields.
In addition to content errors, ChatGPT makes computational and logic errors. ChatGPT
is a language model and not a calculator. Text is generated in response to mathematical
questions with a random number as the outcome, which may not be correct.
Claude and Bard (now Gemini) are two promising new Generative AI writing tools. Like
ChatGPT, they are large language models (LLMs) that can generate text, translate languages,
write different kinds of creative content, and answer questions in an informative
way. However, they have some differences.
Claude is a newer LLM that was released openly to the public in 2023. It is trained on a
dataset of text and code that is specifically curated to be factual and unbiased. This means that Claude is less likely
to generate inaccurate or misleading information than ChatGPT. Claude is also known for its ability to reason and make inferences, which makes
it well-suited for tasks such as solving problems and answering complex questions.
Bard, (now Gemini), released in 2023, is the newest of the three (ChatGPT, Claude, Bard
[Gemini]). It is trained on a dataset of text and code that is larger than the dataset
used to train ChatGPT. Bard (Gemini) is also able to access the real-time web, which
gives it an advantage in terms of the most up-to-date information. Bard (Gemini) is
known for its ability to generate creative text formats, such as poems, code, scripts,
musical pieces, emails, and letters.
The quality of the content generated by ChatGPT and other AI writing tools depends
on the training data and specific prompts provided to the model. There is a risk that
the generated content may contain errors or omissions or lack high quality.
There is a risk that using ChatGPT or other AI Generative models to generate content
may be perceived as plagiarism. It is important to properly cite any materials used
to indicate which parts of the content were generated by a machine.
Models like ChatGPT are only as good as the data they are trained on. There is a risk
that these models could be biased if they are trained on biased data. Take steps to
mitigate it.
Generated content may result in a lack of originality because AI Generative models
are not capable of coming up with new ideas or insights on their own. Content is generated
based on the data it was trained on.
As people use AI Generative, they help train it. Privacy concerns are applicable as
with other places online that require getting an account and sharing information.
By using ChatGPT, users consent to having potentially personal data stored and sold
by OpenAI.
Teaching Strategies
Get familiar with ChatGPT and other Generative AI. Instructors and students might
benefit from learning how to effectively use them. in the teaching and learning process. Doing so includes training on how to use the
system, how to formulate effective questions, and how to interpret responses.
Before using ChatGPT and other Generative AI, it is important to clearly define the
learning outcomes, assessments, instructional content, and activities for the course.
This will help ensure that these tools are used in a way that is aligned with the
overall learning outcomes for the course.
Communicate to students what ChatGPT and other Generative AI are and their limitations.
They are not perfect and may make mistakes or provide incomplete information. It is
important to be clear with students about its limitations.
Students may need guidance on how to use these tools effectively, including how to
phrase questions, how to interpret responses, and how to evaluate the quality of the
information provided. It may be helpful to provide examples of effective and ineffective
uses of ChatGPT and other Generative AI.
These tools can be used in conjunction with other instructional technologies, such
as recorded video lectures and interactive simulations. This can help create a more
engaging and interactive learning experience for students.
Incorporate ChatGPT and other Generative AI as part of a larger instructional strategy. They can be a valuable tool to support learning, users should not rely on it as the
sole means of instruction. It is important to integrate ChatGPT into a broader instructional
strategy that includes other forms of teaching practices, learning activities, and
assessments.
It is important to continuously evaluate and refine Generative AI use to ensure that
it is meeting the learning objectives and to identify areas for improvement. This
can involve analyzing student performance data and soliciting feedback from students
and other instructors.
Emphasize the importance of academic integrity. Foster a culture of academic integrity by encouraging students to take responsibility
for their own academic work, promoting honesty and integrity, and discouraging cheating
and plagiarism.
Monitor student use of ChatGPT and other Generative AI. To ensure that ChatGPT is being used effectively, encourage students to monitor their
use of the tool. Identify any issues or concerns. This can be done through periodic
check-ins, analyzing chat logs, or asking students to reflect on their own use of
ChatGPT.
Monitor and evaluate Generative AI performance. Instructors should regularly monitor and evaluate the accuracy and effectiveness of
ChatGPT in responding to student’s questions and concerns. This will help ensure that
ChatGPT is providing accurate and helpful responses to students.
Authentic assessment: Use authentic assessments that reflect real-world scenarios and require students
to demonstrate their skills and knowledge in practical applications. For example,
you could ask students to design and carry out a scientific experiment or analyze
a real-world case study.
Higher-order thinking: Design assessments that require higher-order thinking skills, such as critical thinking,
problem-solving, and creativity. These types of assessments are less likely to be
replicated by ChatGPT and require students to apply their knowledge and skills in
novel ways.
Collaborative learning: Design assessments that require collaboration between students. Generative AI may
be able to assist with individual tasks, but it cannot replace the benefits of working
in teams and engaging in collaborative learning. For example, you could ask students
to work together on a group project or participate in a debate or discussion.
Feedback: Provide timely and constructive feedback to students on their performance. This can
help students identify areas of strength and weakness and improve their learning outcomes.
Generative AI can provide automated feedback, but it cannot replace the value of personalized
feedback from a human.
By using multiple assessments (i.e., traditional tests, essays, projects, presentations,
and performance-based assessments), instructors can ensure that students are evaluated
on a range of skills and knowledge, making it more difficult for Generative AI to
replicate the assessments.
Get Creative With Your Assignments
Here are 10 ideas for creative assignments adapted for a classroom with chatGPT. Don’t
stop with these — get creative. You can mitigate the risk of students using chatGPT
to cheat, and at the same time improve their knowledge and skills for appropriately
using new AI technologies inside and outside the classroom.
1. Prompt Competition
Identify a major question or challenge in your field or discipline that chatGPT could
write about. Preferably a question with no clear single right answer.
Have students collaborate (in pairs or small teams) on developing 5 to 10 criteria
for assessing chatGPT responses to the major question. For example, chatGPT’s output
references more than one theoretical perspective.
Ask students to individually write a prompt for chatGPT to answer the major question.
Have students use their criteria to judge the responses of other students (in the
pair or small team), and rate the chatGPT prompts/responses from best to worst.
2. Reflect and Improve
Ask students to individually identify a major question or challenge in your field
or discipline that chatGPT could write about.
Have students use chatGPT to write a response to their question or challenge.
Ask students to reflect on chatGPT’s output (e.g., what is correct, incorrect, what
they don’t know if it is correct or incorrect, what should they look up elsewhere
to verify, what should they ask chatGPT next).
UsingTrack Changesin MS Word orSuggestingin Google Docs, have students improve the output of chatGPT (e.g., correcting errors
or misinformation, expanding on shallow content).
Have students submit their prompt and the improved chatGPT response with their added
content highlighted.
3. Re-vision
Ask students to individually identify a major question or challenge in your field
or discipline that chatGPT could write about.
Have students use chatGPT to write a response to their question or challenge.
George Heardis attributed with saying “The true meaning of the word revision is this: to see again.”
Have students revise (write again) chatGPT’s output from a different angel. For instance,
take a different perspective, apply a critical lens, expand on a particular concept,
or correct aspects of the output that could cause their peers to misunderstand or
misinterpret.
4. Dual Assignments
Give students a choice between two versions of the same assignment. One version for
those that want to use chatGPT and one for those who don’t.
For those who choose to use chatGPT, they have to submit their prompt(s) and the chatGPT
output. UsingTrack Changesin MS Word orSuggestingin Google Docs, have students add depth, clarify misinformation, offer alternative
perspectives, and make other improvements to the chatGPT output.
For those who choose to complete the assignment without chatGPT, they should complete
the assignment and sign a statement that chatGPT was not used.
Grade both assignments on how well students illustrate their depth of knowledge through
either (a) their changes to chatGPT’s output, or (b) their original writing.
5. Mind Maps
Since chatGPT can’t natively make visual representations of content (see note below),
have students create mind maps (aka, associative maps, spider map, process maps) to
illustrate the connections between ideas, concepts, approaches, or theories in your
field or discipline.
The more details or levels that students add to their mind minds, the easier it will
be for them to demonstrate their newly acquired knowledge and skills.
6. Debates
Have students debate a major question or challenge in your field or discipline. Even
short debates can deepen learning and get students to look at topics from varied perspectives.
You can choose if students are allowed to use chatGPT in their preparation for the
debate’s opening statements.
Debates can be done in differentformats, and the length of times for speeches can vary depending on how much time and how
many students are in your course.
7. Videos or Podcasts
Rather than written essays, have students make videos or audio recordings as the medium
for sharing their knowledge.
Using a video-based tool (such as VoiceThread, FlipGrid, or Zoom) can make the process
easier for students.
Students can also record audio podcasts on their phone or computer if visuals are
not required for the content of the assignment.
8. Explain Your Thinking
Give the assignment as usual, but in addition require that students use UsingTrack Changesin MS Word orSuggestingin Google Docs to explain at least 8 to 10 steps of their thinking as comments added
to the text.
Students can describe, for instance, the steps in their logic, their problem solving
or writing process, or the development of their theoretical path.
Students could also document their thinking with audio recordings or videos.
9. 2x2 Matrix
Have students create a 2x2 matrix relating two concepts covered in the course. For
instance, what are shared and different defining characteristics of concepts or processes.
A simpler version of this assignment is to have students develop Venn Diagrams for
comparing important concepts or processes.
10. Next Time
Ask students to use chatGPT to answer an essay question about a major question or
challenge in your field or discipline.
Have students reflect on their learning about the topic based on using chatGPT, and
to write down 5 things they learned about the topic from chatGPT.
Have students design a new assignment that doesn’t allow for the use of chatGPT but
that would allow them (or other students) to demonstrate their learning. For example,
they might suggest a group project, or mind map assignment.
Instructors should educate students on the importance of academic honesty and integrity.
They should explain to students that using AI-powered tools like ChatGPT is not a
substitute for their critical thinking and analysis and that they must ensure that
their work is original and properly cited.
Instructors should provide students with clear instructions on how to use Generative
AI in their assignments. This includes guidelines on how to cite the tool as a source
and how to integrate its output into their work.
Instructors should encourage students to think critically when using ChatGPT. They
should explain that the tool is only as good as the input it receives and that students
must carefully evaluate the relevance and accuracy of the tool's output.
Instructors should monitor students' use of Generative AI to ensure that they are
using the tool ethically and responsibly. They should look for signs of plagiarism
or other forms of academic misconduct.
Instructors should use multiple assessment methods to evaluate students' work. This
can help reduce the temptation for students to rely solely on Generative AI to complete
their assignments.
Instructors should provide feedback on student work that highlights areas where they
can improve their use of Generative AI. This can include suggestions for how to better
integrate the tool's output into their work or how to improve the relevance and accuracy
of the tool's output.
Research Tool: Generative AI can be used as a research tool to help students explore different topics
and find answers to their questions. Instructors can encourage students to use these
tools to research topics related to their coursework or assignments and provide guidance
on how to use the tool effectively.
Writing Assistance: Generative AI serves as a writing assistant to help students generate ideas and improve their writing
skills. Instructors can encourage students to use these tools to generate topic ideas,
brainstorm content, and get feedback on their writing.
Language Practice: Generative AI as a language practice tool helps students improve their language skills. Instructors
can encourage students to practice their writing or speaking skills by engaging in
conversations in the target language.
Interactive Learning: Generative AI can be used as an interactive learning tool to engage students in interactive
and personalized learning experiences. Instructors can design interactive learning
activities using ChatGPT, such as quizzes, games, or simulations.
Customized Content: Generative AI can be used to create customized content tailored to individual student
needs. Instructors can use ChatGPT to generate personalized study materials for students,
such as summaries of course content, flashcards, or study guides.
Research Assignment: In this assignment, students are required to use Generative AI to research a topic
related to their coursework or assignment. Students will be required to submit a written
report on their findings, including a summary of the information they found, the sources
they used, and their evaluation of the quality and reliability of the information.
Grading criteria could include the quality of the report, the relevance and accuracy
of the information found, and the evaluation of the sources used.
Writing Activity: In this activity, students will use Generative AI to generate ideas and improve their
writing skills. Students will be given a writing prompt, and they will use ChatGPT
to generate ideas and structure their writing. They will then write a draft of their
essay, using the ideas and structure provided by ChatGPT. Grading criteria could include
the coherence and organization of the essay, the quality of the ideas generated, and
the relevance of the content to the writing prompt.
Language Practice: In this activity, students will practice their language skills by engaging
in a conversation in the target language. Students will be required to submit a transcript
of their conversation, along with a reflection on their language proficiency and areas
for improvement. Grading criteria could include the accuracy and fluency of the student's
language use, the complexity of the conversation, and the reflection on their language
skills.
Interactive Learning: In this activity, students will engage in an interactive learning experience using
Generative AI. Students will be given a set of questions related to the course content,
and they will use Generative AI to find the answers. They will then submit their answers
along with a reflection on their learning experience. Grading criteria could include
the accuracy of the answers, the quality of the reflection, and the engagement and
participation in the activity.
Customized Content: In this assignment, students will use Generative AI to generate customized study
materials tailored to their individual needs. Students will identify areas where they
need additional support or clarification, and they will use ChatGPT to generate study
materials such as summaries, flashcards, or study guides. They will then use these
materials to prepare for an exam or assignment. Grading criteria could include the
quality and relevance of the study materials, the effectiveness of the materials in
helping the student prepare, and the reflection on their learning experience.