Skip to Content

Center for Teaching Excellence

  • Generative Artificial Intelligence (GenAI)

Generative Artificial Intelligence (GenAI)

In November 2022, OpenAI released ChatGPT (Chat Generative Pre-Trained Transformer), which prompted conversations among educators about its potential to impact the higher education landscape. Just a few months later, Anthropic released Claude to directly compete with ChatGPT, while Google unveiled Bard (now Gemini) as its own AI chatbot. This USC 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.

Learning Strategies

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.


Carvalho, L., Martinez-Maldonado, R., Tsai, Y.-S., Markauskaite, L., & De Laat, M. (2022). How can we design for learning in an AI world? Computers and Education: Artificial Intelligence, 3

Chavez, M. R., Butler, T. S., Rekawek, P., Heo, H., & Kinzler, W. L. (2023). ChatGPT (Generative Pre-trained Transformer): Why we should embrace this technology. American Journal of Obstetrics and Gynecology

Choi, E. P. H., Lee, J. J., Ho, M-H., Kwok, J. Y. Y., & Lok, K. Y. W. (2023). Chatting or cheating? The impact of ChatGPT and other artificial intelligence language models on nurse education. Nurse Education Today, 125

Claude. (n.d.). Getting started with Claude

Eke, D. O. (2023). ChatGPT and the rise of Generative AI: Threat to academic integrity? Journal of Responsible Technology, 13

Google. (n.d.). What can Bard do and other frequently asked questions? Bard is now Gemini.

Iskender, A. (2023). Holy or Unholy? Interview with Open AI’s ChatGPT. European Journal of Tourism Research, 34, 1–11. 

Mollick, E. R., & Mollick, L. (2022). New modes of learning enabled by AI Chatbots: Three methods and assignments (2022). SSRN Electronic Journal

OpenAI. (2022, September 2). About OpenAI. OpenAI. Retrieved January 24, 2023.

Pavlik, J. V. (2023). Collaborating with ChatGPT: Considering the implications of Generative Artificial Intelligence for Journalism and Media Education. Journalism & Mass Communication Educator, 78(1), 84–93. 

Perkins, M. (2023). Academic Integrity considerations of AI large language models in the post-pandemic era: ChatGPT and beyond. Journal of University Teaching & Learning Practice, 20(2), 1–24. 

Tlili, A., Shehata, B., Adarkwah, M. A., Bozkurt, A., Hickey, D. T., Huang, R., & Agyemang, B. (2023). What if the devil is my guardian angel: ChatGPT as a case study of using chatbots in education. Smart Learning Environments, 10(1), 1–24.

Challenge the conventional. Create the exceptional. No Limits.