Overview
Though AI has already changed academic work and postsecondary education, it has not done so in the same ways or to the same extent within departments or across disciplines and divisions. No uniform consensus has emerged either at Rhodes or within academia regarding its place in higher education. As attitudes toward AI differ significantly across the college, the integration of artificial intelligence into teaching, scholarship, and creative work is left to the discretion of each faculty member. Faculty have the autonomy to design and implement classroom experiences in accordance with the goals of their courses and the subjects they teach, but you are encouraged to consult with your disciplinary colleagues to be sure that your students are gaining the full range of skills they need to succeed in your field.
The Rhodes College catalog states that “a student must not adopt or reproduce ideas, words, or statements that are not their own without appropriate acknowledgment. This prohibition extends to the output of generative Artificial Intelligence (AI) tools and editors, including, but not limited to, text, image, sound, video, coding content, and online translators. Use of AI-generated content in the completion of coursework without proper acknowledgment is considered an act of plagiarism unless such use is expressly permitted by the course’s instructor.” Because permissions may vary from one class or department to another, it is crucial to communicate clearly with students regarding the use or restriction of generative AI. The following pages include a range of syllabus language that the AI Task Force has standardized for use across the college. We recommend that you adopt or adapt some of this language for your own course.
What is Generative AI?
Generative AI (GAI) uses generative models to learn the patterns and structures latent in their training data and produce outputs similar to human-created text, images, video, music, etc. GAIs rely on massive data sets and complex mathematical models, but users can often interact with them through natural language prompts.
Popular AI Tools and Applications
AI tools can operate as stand-alone websites or apps or as agents within other platforms and programs.
- ChatGPT (OpenAI), Claude (Anthropic), Grok (xAI), Gemini (Google), Perplexity AI (Perplexity), and DeepSeek (DeepSeek) are just a few of the most popular chatbots and search engines powered by Large Language Models (LLMs).
- Retrieval Augmented Generation (RAG) is a subset of the LLM category that builds extra guardrails to minimize hallucination. RAGs either limit their content reference to or prioritize files that you upload, but they are still generating text using the underlying LLM statistical models. So hallucination is still possible, but is reduced. Two popular RAGs for education are NotebookLM (Google) and Teachanything.ai (George Washington University - uses open access AI models).
- Popular text-to-image diffusion model GAIs include DALL-E (OpenAI), Midjourney (Midjourney), and Stable Diffusion (Stability AI). They are not text-based like LLMs but rather predict pixels.
- Free text-to-video GAIs are falling quickly, but some people use Kling, Dream Machine, or ComfyUI.
- Agentic AI is a growing field as of 2026 and includes AI tools that receive a high-level human instruction and independently carry out tasks and manage files. Popular agentic AIs include ClaudeCode (Anthropic), Comet browser (Perplexity), and tools such as the short-lived Einstein AI based on the OpenClaw open source code framework.
Microsoft Copilot: Rhodes faculty, staff, and students have access to Copilot, a GAI chatbot based on OpenAI's GPT-4 and GPT-5 series of large language models. Our institutional license provides Copilot users with data protections that "free" or premium subscription chatbots do not. Copilot can be accessed directly through Outlook.
Potential Benefits and Pertinent Warnings
AI remains a young field, and it has since its inception inspired the same anxieties, hopes, and hype that often attend the rise of any new technology. Some view GAI as a revolutionary and democratizing technology that will accelerate the production of new knowledge and eliminate the more tedious aspects of modern work and life; others consider it a Trojan Horse for an extractive economics based on leveraged data-brokering that does not deserve the name "intelligence." Two things can be true, so users should approach or avoid AI critically, conscientiously, and in accordance with the expectations of the college and the values of a liberal arts education.
Faculty and staff at Rhodes and elsewhere have reported using GAI to:
- Generate case events and distractors for multiple-choice quizzes and exams: GAIs can quickly iterate on existing questions with different examples, numbers, etc., which can be useful for offering different quiz and exam options or practice tests, so long as the faculty member checks them for accuracy
- Provide outputs against which students can test their knowledge and skills (fact-checking, analysis, etc.): some studies have shown that AI tutoring is highly effective, when carefully designed and monitored by instructors, and it's a use to which students are already putting GAIs on their own
- Assist in writing letters of recommendation: such letters are a nationwide area of debate, but it is possible to write such a letter effectively. If you want to use GAI in this way, please make sure that you are doing what is best for the student - which might include declining to write a recommendation or telling them that you plan to use GAI
- Assist in writing administrative documents/reports: if you choose to do so, please ensure that you are using a FERPA-compliant version of GAI (such as the provided CoPilot) and do not include any sensitive or identifiable student information
- Take minutes in departmental and committee meetings: same as above, and be extra cautious because a listening GAI may hear things that you hadn't intended to include in the minutes
- Assist in writing code: GAI can produce computer code based on prompting, and so-called "vibe coding" is a form of agentic AI that can produce code, including functional websites, based purely on natural language instructions
- Generate or supplement images, music, video, powerpoints, mindmaps: RAGs can provide ways for students to access the same content in multiple formats, increasing their engagement
- Brainstorm exercises, assignments, rubrics, syllabuses: Caveat emptor, but with appropriate editing, some reporting saving time and having more customized materials for their classes
They have also raised concerns about GAI regarding:
- Academic Dishonesty and Plagiarism: GAIs make it easier for students to generate formal essays and assignments as well as casual written responses and even in-class comments in seconds; these can be tailored to reflect specified educational levels, vocabularies, and styles and to include grammatical/syntactical errors suggestive of authenticity
- AI Detection Programs: despite the claims of their purveyors, AI detection remains unreliable at best and can lead to false accusations or suspicions that undermine mutual trust
- Deskilling: some studies indicate that cognitive offloading to GAI tools may degrade critical thinking skills, reading comprehension, and recall
- Misinformation and Disinformation: GAI systems can generate inaccurate, outdated, or manipulated outputs. OpenAI has admitted that LLM "hallucinations" are mathematically inevitable; reporting suggests that xAI has “tweaked” Grok to align with a particular political ideology in response to certain queries; DeepSeek is likely also censored at the training and application levels; and Sora 2 can produce fairly convincing video deepfakes.
- Environmental Impact: training GAIs and operating data centers requires substantial amounts of energy, water, and land, and produces significant carbon emissions and e-waste (n.b. Colossus, an xAI supercomputing facility, is located in South Memphis).
- Copyright and Intellectual Property: LLMs are trained on vast amounts of data--text, images, audio, and video--much of which may have been acquired and used without proper license, attribution, or compensation. Legislation lags behind technological developments, and lawsuits are ongoing.
Neither of these lists is exhaustive, and as new tools (or updated versions) become available, practical use cases and ethical concerns surrounding them will continue to emerge. What is most important, however, is that you are still responsible for anything you produce, with or without GAI. You need to vouch for the accuracy of and stand behind the content of anything you present to students, submit to
Reference Guides and Tutorials