As the modern tech landscape continues to invest in the development of AI tools such as a large language models, their presence on CNU campus becomes more and more likely for students and faculty alike. In particular, with the development of job opportunities centered around the usage of AI at companies as large as Microsoft, knowing how to navigate such usage in a safe and ethical manner becomes increasingly important. With this in mind, CNU's IT department has developed a set of recommendations for implementing the use of AI tools on campus:
Potential Issues
What IS an AI?
Transparency and Oversight
Unintended Biases and Discrimination
A strong general rule when judging an AI model is that the model cannot create anything it hasn't already seen before–meaning that the output of an AI will always be the result of all the data that was input into the model to train it. This can result in some key issues in the output of large AI models going completely unnoticed by users who assume the model is more intelligent than it is. One example is that a model trained on biased data will replicate that bias in its results; for example, if a medical AI used to diagnose skin diseases is trained using only data from white patients, it may struggle to properly identify medical issues on bodies of color, and result in fewer correct diagnoses on nonwhite skin as a result. However, the model will be seen as presenting its output as purely objective data. Uncritical usage of models fed on biased data can feed into deeply entrenched systems of discrimination, and furthermore mask that discrimination by presenting it as the unbiased truth from a model unaffected by personal bigotry.
Misinformation
Environmental Impact
Educational Shortcutting
A common concern stated by educators regarding AI is that the use of AI models will enable students to hide a lack of knowledge about a subject they were intended to learn by using an AI model that can generate the knowledge needed to pass the class. For example, a student who does not understand a unit about Shakespeare, instead of demonstrating his own lack of mastery by writing an essay that would not be received well, may use an AI language model to generate an essay that would demonstrate knowledge he has not retained. Such 'shortcuts', if not properly managed, retain the risk of making students dependent on tech tools they do not understand rather than learning the knowledge and skills they need for themselves, which can have dangerous consequences if said students enter fields where said knowledge is required to perform a job properly and safely.
Training Data, Privacy and Data Governance
For Faculty
Create Clear Expectations
CNU does not currently have specific rules in their Academic Integrity policy regarding the use of AI tools in the classroom. As a result, it is the responsibility of the faculty to outline their expectations for their classroom in clear terms. These terms may vary depending on your curriculum, your personal teaching philosophy, or whether you intend to incorporate such tools into the classroom yourself for specific assignments. This provides you the freedom to state whether or not you consider the usage of AI tools to be a violation of academic integrity or not, based on the learning expectations you have for your students. It is currently recommended that faculty outline these expectations in their syllabus, as well as clearly state them for students at the beginning of the semester. Failure to do so may create ambiguity in the face of a potential incident.
Scaffold Usage to Encourage Learning
Common worries about AI use in the classroom include the possibility of it providing shortcuts for students that prevent them from learning. However, this is not a concern unique to AI tools; similar worries have followed almost every new introduction of technology into education. Similarly to tech like calculators and LMS platforms like Blackboard, AI can create valuable learning opportunities when implemented carefully. Faculty considering the usage of AI tools in the classroom should carefully consider what kinds of assignments can benefit from using AI, what students are intended to learn from these assignments, and how such usage can encourage learning rather than provide shortcuts. A recent example may include the use of AI tools in a writing class to generate text that students then edit and critique. Rather than bypassing the students' need to learn writing for themselves, this usage can enable students to learn how to edit and critique existing writing by providing a neutral sample for them to practice on.
If you are struggling to find an appropriate way to scaffold AI tools in your curriculum, please reach out to Academic Technologies, or the Center for Effective Teaching, for assistance.
Provide Materials on Smart and Ethical Usage
Students unfamiliar with AI tools may come into your classroom either never having used an AI tool, or already using these tools in ways that run counter to their best interest. A common example includes a trend of incoming students using ChatGPT as a Google Search analog, unaware that the AI model's algorithm does not prioritize the accuracy of information being output. As a result, we recommend that alongside your expectations for academic integrity, faculty provide educational materials on the nature and safe/ethical usage of AI, especially if the usage of AI tools has been scaffolded into part of your curriculum. You can use the material in this document as a starting point, but can also use recent news articles, links to current models' training data, or other documents.
Do Not Depend on Detection Software
At this time, there are no tech tools that can reliably detect when a piece of text was generated using an AI tool. As a result, using tools that can claim to do so in order to check student work can not be currently recommended. We instead recommend that faculty use more analog methods to investigate student work, such as checking the accuracy of the text and noting sudden stylistic changes. Depending on the context and submission method of the assignment, IT department may investigate text submitted into Scholar to provide further evidence, but this may also be inconclusive.
Carefully Vet Suggested Tools
For Students
Check Classroom Guidelines
Before any other consideration, a student should confirm with their instructor whether the usage of AI tools in their classroom is approved or not. Different instructors may have widely varying policies depending on the type of class they teach and their personal teaching philosophy. As a result, what may be considered a valuable and useful tool in one classroom may be considered a breach of academic integrity, plagiarism, or outright cheating in another. Other faculty may have a more nuanced take, encouraging their use in some contexts and discouraging them in others. Faculty are being encouraged to detail their specific expectations in their course syllabus, which will be an ideal place for you to first check; if said expectations are not present or unclear, err on the side of caution by checking with the instructor directly before using any AI assistance.
Verify All Facts And Sources
The coding behind AI tools such as large language models does not differentiate between true and false information, and has often been seen generating false facts or nonexistent references simply by mimicking the structure of the texts they were trained on. In addition, the use of inaccurate, misleading, or even satirical content in a model's training data may also result in the model returning similar inaccuracies. As a result, it is recommended that any user of AI tools carefully review the output of said tool for accuracy. Flag any false or nonsense information produced by the model, as well as any nonexistent sources the text may claim to be utilizing, and use your own knowledge to correct them.
Use Ethically-Trained Models
AI models are trained on large sets of outside data to develop their algorithms, which can come from a variety of places depending on the type of model. This can be the source of multiple kinds of unethical and even illegal behavior on the part of the model, including the generation of unintended biases, copyright violation, and outright theft of confidential data such as medical records. Work is being done by multiple organizations to enforce transparency regarding the sourcing of training data used to train popular AI models such as ChatGPT; this work has included the filing of multiple lawsuits against models accused of scraping data from unethical or illegal sources. In order to ensure that student work does not unintentionally replicate these past mistakes, we recommend that all users research the intended AI model's history with their training data, and avoid using models that do not provide information about where said data is sourced from.
Edit, Edit, Edit
Another common failing of AI generated models is their limited ability to replicate tone or style. While large language models can reliably mimic common writing styles, or the style of well-known individuals whose work may be part of their training data, it is not guaranteed that the resulting output could mimic your own way of writing or speaking reliably. In addition to providing opportunities to correct any misinformation, heavily editing the output of an AI tool allows you to customize the work so it more closely aligns with your body of work as a whole.
Practice Full Disclosure
To untrained eyes, the results of an AI tool like ChatGPT can be indistinguishable from original work generated by a human mind. However, this can create issues of academic integrity, particularly in situations where the implied expectation is that any writing used in a classroom setting is human-generated. When in situations where the expectations are not clearly outlined, it is suggested that a disclaimer be attached to any AI-assisted work notifying anyone reading or reviewing the work that such assistance is present. This can bypass potential confusion and misattribution of AI-presented work, and can further clarify the cause of other potential issues that may arise as a result, such as those listed above.