Generative Artificial Intelligence for Teaching, Research and Learning

Below you will find a brief overview of the use cases for generative AI in academic settings. This list is not comprehensive and the tools mentioned are intended as examples rather than endorsements/recommendations. UC Davis Libraries does not license these tools or provide user support for them.

Chatbots like ChatGPT can be used to "rapidly generate human-like texts with a unique capability to conduct a wide variety of language prompts, including question-answering, translation, summarization, and text generation." (Bin-Nashwan, S.A., et al, 2023)

Popular tools include: ChatGPT (OpenAI), Gemini (Google), Claude, and Copilot (Microsoft). UC Davis licenses Microsoft Copilot. This tool also can generate images.

These use natural language text prompts to generate images and videos to aid users in the visualization of ideas, creation of illustrations, visual aids, production of graphic designs, and the modification of images, among others. Popular tools include DALL-E (OpenAI) and Midjourney. 

These can help users enhance research output through AI driven literature search, review, and synthesis and pull from caches of academic articles such as Semantic Scholar and Cochrane. Many of these tools are in early stages of development and cannot replace the precision and breadth of a traditional literature search. Popular tools include: Elicit (Natural language literature searching,) Scite (citation search and analysis,) and Consensus (literature search and synthesis.)

• Whitfield, Sharon, and Melissa A. Hofmann. "Elicit: AI Literature Review Research Assistant." Public Services Quarterly 19.3 (2023): 201-207.

In addition to tools like CoPilot and ChatGPT, the following can enhance digital accessibility for users by generating captions, summaries, transcriptions, and translations of text, audio, and video. For example, VidBy can generate captions and subtitles for videos, Otter.ai and AudioPen.ai are speech to text transcription tools, and Wiseone.io and Humata.ai can summarize and paraphrase PDFs. Another mode of accessibility these tools can be used for are to distribute research to a wider readership. For example, Schmitz, 2023 states: "Artificial intelligence language models have the potential to generate lay abstracts that are consistent and accurate while reducing the potential for misinterpretation or bias."  

There are a growing number of tools available to help users write code in a variety of programming languages. Many of these tools can take a plain-language description of what you would like the code to do and then write a basic, generic block of code that accomplishes that task. These blocks can be pieced together by the user to build bigger scripts and programs. In additional to ChatGPT, popular tools include GitHub Copilot and OpenAI Codex.