Skip to Main Content
Link to CSU Home
Link to CSU Home
Students in the Library

Michael Schwartz Library

216.687.2478
Students in the Connection Lounge
Michael Schwartz Library

Generative AI Tools and Information Literacy

What is generative AI?

Generative artificial intelligence (AI) consists of deep learning models that generate content like text, images, and video based on the data that it was trained on. Unlike a search engine, generative AI doesn't search through lots of human-generated content to provide you an answer. Instead, it uses mathematical models to predict what word statistically should go next. Generative AI tools are changing at a rapid pace, improving their outputs, and being used to develop a wide variety of tools to make human work easier and more effective. While these tools are not always accurate and produce what are called hallucinations (made up or biased content that isn't accurate), many people have found generative AI tools useful in their work and everyday life. 

Types of Generative AI Tools

This type of generative AI uses massive amounts of text from books, articles, and websites to determine the common relationships between terms and concepts in human language. It then uses these calculations to generate new text. It can produce a wide range of textual content, from reports, to essays, to even poetry. 

This type of generative AI uses datasets that consist of images with captions or descriptive text to "understand" human concepts and their relationships. It can then output images that reflect these concepts, including novel combinations of concepts in visual form. For example, when asked to produce an image of a ladybug as a computer mouse, it knows both concepts and can combine them into an image. 

A Giant Ladybug - AI Generated Artwork - NightCafe Creator

This type of generative AI uses music tracks along with metadata (information about the music), along with music lyrics to detect trends and patterns in human-generated music. It then can use this information to generate new musical works.

Prompts: Prompt engineering for text to music models might involve type of instrumentation, genre, specific references to artists, mood and tone, as well as more musical prompts, like rhythm or tempo, key, etc. 

Some models allow additional controls, like uploading a simple melody to incorporate as well as supplying lyrics to be sung, rapped, or spoken.

Stem Splitting Tools

One of the most popular uses of generative AI is to help with coding projects. Since coding is like writing in another language, generative AI is able to ingest coding languages and learn the relationships and functions of these languages. The AI tool can then be used by coders to more efficiently write and correct code. 

Because videos involve visuals, audio, and, sometimes, text/voice, there are multiple ways that generative AI tools can create videos. Some use existing videos as their dataset to generate new videos; others use audio, visual, and text sources to come together into a new video; and others have been trained how to use a video editing software, so they can improve a video you have created in that software, such as by adding captions or animations.  

Tools for Getting Started

ChatGPT is the most common tool from OpenAI so far. It requires users to create a login, but it's free. There is a paid version called ChatGPT Plus, which offers priority access to new features. 

  • Conversational, natural-language interface - it responds to prompts/questions in complete sentences rather than retrieving and listing search results.
  • Well thought-out prompts produce more useful results.
  • Data training cutoff was September 2021, so it has little to no information about events since then.
  • Prone to errors and "hallucinations." 

Stable Diffusion is a free and easy-to-use text-to image generator. 

DALL-E 2 creates images and art from a text description entered by a user. Users may also upload an image to edit, or to create variations inspired by the original. It is currently not free, but it's inexpensive

  • A credit is charged for each generated image. The minimum cost is $15 for 115 credits. 
  • Early adopters who signed up before April 6, 2023 receive free credits that are replenished monthly.
  • Policy to prohibit harm: https://labs.openai.com/policies/content-policy.