Additional Information

Generative AI & Large Language Models

genAI at USF

Generative Artificial Intelligence provides opportunities to ethically and effectively advance our institutional mission. USF is committed to actively engaging with this rapidly developing technology to maximize these opportunities.   

Visit USF's generative AI website for comprehensive guidance on ethical usage, syllabus and course policy recommendations, citations, AI events, and more. 

Use of Generative AI and Large Language Models in ETDs

The Office of Graduate Studies maintains the University's focus on academic integrity and upholds the goal of protecting students who publish their electronic theses and dissertations through USF. The use of generative AI and Large Language Models (LLMs) can pose significant challenges in regard to both academic integrity and the integrity of the work due to “hallucinations” (because LLMs create content by predicting the most likely next word in a sequence, they sometimes generate content that is not accurate or content that is biased). 

OGS Guidelines are based on USF’s AI Policy:

“Per USF Policy, use of any tool or assistive technology for completion of coursework must be authorized by your instructor first. If you choose to use ChatGPT or some other generative AI technology for writing, be sure you are transparent about your use of it with your teachers. Each citation style has different recommendations for citing generative AI tools like ChatGPT.“   

Source

A student who wishes to use generative AI or LLMs in the preparation of their ETD must:

  1. Obtain permission from their entire committee
  2. Be transparent about AI use and follow Citation Guidelines
  3. Clearly Cite the Model and Version Used
  4. Be vigilant in reviewing and correcting all “hallucinations”
  5. Accept and acknowledge that they, the student author, are ultimately responsible for any inaccurate or biased information in their ETD even if it was generated by AI
  6. Obtain Copyright Permission for any images or content created by others and then run through Generative AI

Citing AI-Generated Content

Transparency and honesty are an important part of maintaining academic integrity. 

Citing the use of any generative AI for written content and images is required in your ETD, and generative AI can only be used in your ETD with permission from the entire committee. 

If permission is obtained, use the following guidelines to cite.

The citation must include:

  • AI Program,
  • Version,
  • Directives/prompt,
  • Date of access  

Each style has different recommendations for how to cite generative AI tools, like Chat GPT seen below. 

APA 7: 

Open AI.  (Year). ChatGPT (month day version) [Large Language Model].
          https://chat.openai.com/chat 

MLA 9:  

"Prompt text" prompt.  ChatGPT, day month. version, OpenAI, day month year,
          chat.openai.com/chat

Chicago: 

ChatGPT, response to "Prompt text," OpenAI, month, day, year,
          https://chat.openai.com/chat 

Read the USF Library’s page on Citing Generative AI tools for more details. 

Please consider other publisher’s policies on AI as well. 

Source

Generative AI programs include, but are not limited to:

  • ChatGPT, GPT-3, GPT-3.5, GPT-4
  • BERT
  • Claude
  • Falcon40B
  • Lamda
  • Llama
  • Orca
  • Palm
  • Phi-1
  • StableLM
  • Vicuna33b
  • AlphaCode
  • GitHub Copilot
  • DALL-E 2
  • StyleGAN
  • Bardeen
  • Copy.ai
  • 14Descript
  • Type Studio
  • Designs.ai

NOTE: Plagiarism discovered because of AI-generated ETD material, images and text, is subject to an academic integrity review and sanctions.