Skip to Main Content
Today's Hours

Generative Artificial Intelligence

This guide provides information for VCU students, faculty and staff on the topic of generative artificial intelligence tools so that they may assess practical and ethical issues relevant to their work within an academic setting.

Three Categories of Generative AI

There are hundreds of GenAI tools (and more every day). For one of the most comprehensive (and easily searchable) lists, we recommend checking out the website "there's an AI for that". However, if the sheer amount of tools is overwhelming, a good starting point can be to think of them as having three broad categories of outputs: text, multimedia, and code.

text iconText

Since most GenAI generators already use text as the input, it makes sense that the vast majority of these tools generate text as the output as well. This type of artificial intelligence is often called a Large Language Model (LLM), which is an algorithm that uses deep learning (neural networks) to process huge amounts of data (mostly scraped from the internet). The result is a system that can produce natural language responding to prompts, such as answers to questions. The most well-known of these generators is ChatGPT, which is OpenAI's large language model. 

Text Generating Examples:

Tool: ChatGPT

Prompt: Write a pop song about generative AI.

(Verse 1) In a world of wires and code, A new kind of magic unfolds, Generative AI, the future's bright, Creating wonders day and night.

(Pre-Chorus) Lines of code dance in the air, Dreams and visions everywhere, From pixels to melodies, it's no surprise, Generative AI paints the skies.

(Chorus) Oh, Generative AI, you're the melody of tomorrow, Painting colors we've never seen, erasing sorrow, With algorithms and data, you're making us believe, In a world where imagination and tech can truly achieve.

multimedia iconMultimedia

GenAI can also be used to generate multimedia objects from text, such as images, videos, vocals, music, and even 3D objects! In these cases, deep learning attempts to find patterns within a specific type of media (images, sounds, etc.) and uses this data to synthesize new media based on prompts from the user. 

Multimedia Generating Example:

Tools: DALL-E (image synthesizer), HeyGen (video synthesizer), and VoiceMod (music synthesizer) 

Steps: 

  • Images were generated in DALL-E with the prompt “a photo of ram dressed like a librarian, wearing a bowtie and glasses, background full of books”
  • Music was generated in VoiceMod, using the above lyrics from ChatGPT. Three different vocals were synthesized. 
  • These images were then animated in HeyGen by combining them with the singing vocals. 
  • All three synthetic videos were then combined using Adobe Premiere Pro to sync them to play together. 

 

code iconCode

One of the most promising applications of GenAI, however, is coding itself! There are many AI tools that can help all kinds of levels of coders. For example, a beginner can use these tools to generate some simple code from scratch, or an advanced programmer can collaborate with an AI assistant to help you find errors and streamline your code. 

Code Generating Examples:

Tool: JIT.codes

Prompt: Create a looping animation of a circle changing colors and scale, with text on top "Rambot"

See the Pen RAMBOT by okkeyes (@okkeyes) on CodePen.