University Center for Teaching and Learning

Generative artificial intelligence (AI) tools use machine learning models trained on massive pools of information to learn patterns from data to create novel content like text, images, audio, or video in response to a prompt. Unlike internet searches, generative AI tools do not use algorithms to locate and curate existing sources. Instead, they create new content by predicting what word, sound, or pixel would come next in a pattern.

Learn more about the history of generative AI in education and terms associated with AI. Learn more about how generative AI functions.

What are the capabilities and limitations of generative AI tools?

Generative AI tools can:

  • Respond to prompts and questions
  • Summarize and synthesize information
  • Revise and edit content
  • Generate creative works like musical compositions, stories, jokes, and poems
  • Write and correct code
  • Manipulate data
  • Play games

In November 2023, ChatGPT released custom GPTs that can be trained to perform more specific tasks. For example, one faculty member in Pitt’s Psychology Department created an extra credit assignment in a research methods course that involved students creating a custom GPT to check and help them correct APA formatting issues in their own assignments. Custom GPTs can be used for a range of tasks like tutoring students on specific subjects, skills, or assignments.

Generative AI tools are constantly learning and evolving but, as of the date of this publication, some limitations include:

  • With some generative AI tools, consistently integrating real research into text is at best a weak functionality. Some AI tools, for example, can produce text with a reference list, but the references often do not correspond to the text created or are fake citations made of a mix of real publication information from multiple sources. More-recently released tools like GPT4 and Bing Copilot have more-sophisticated research integration capabilities.
  • ChatGPT 3.5 (the free version of ChatGPT) is trained using data available up until January 2022. GPT-4 is trained using data available up until April 2023 and now has plug-ins that provide some internet connectivity. Other tools, such as Bard and Bing Copilot, are always internet connected and have access to current information.
  • Generative AI can still compose potentially incorrect, oversimplified, unsophisticated, or biased responses to questions or prompts.

Although many faculty are understandably concerned about the academic integrity implications of students using generative AI tools like ChatGPT, generative AI can also enhance and support your teaching.

Generative AI Tools

This list is not comprehensive but features some of the most widely used generative AI tools. Tools with free versions are indicated with asterisks.

To request that we add a tool to these lists, contact us at teaching@pitt.edu.

Chatbots and Text Generators

  • Bing Chat*
  • ChatGPT*
  • Claude
  • Copilot*
  • Google Bard*
  • GPT-4
  • Poe

Research Tools

  • Connected Papers
  • Consensus
  • Elicit: The AI Research Assistant
  • JSTOR Text Analyzer
  • Litmaps
  • Open Knowledge Maps
  • Research Rabbit
  • Semantic Scholar

Image Generators

  • DALL-E2
  • Midjourney

Audio Generators

  • Play.ht* (voice)
  • Murf.AI* (voice)
  • Mubert* (music)
  • Soundful (music)

Video Generators

  • Synthesia

Code Generators

  • GitHub Copilot
  • Codacy

Research Tools

  • Elicit (summarizes and synthesizes sources for literature reviews)
  • Discuss Genie (qualitative research AI assistant)