Generative AI Guide

Images generated using Bing Copilot, with the following prompts: (1) "A student using a computer at the library as abstract art", (2) "A student using a computer at the library in an oil painting style", (3) "A student using a computer at the library", (4) "A student using a computer at the library in a cartoon style"

Academic Integrity

Consult with your instructor before using any form of artificial intelligence (ChatGPT, Grammarly, etc.) for your work. If you have been given permission to use AI tools, you must always cite the tool you used.

Under Policy 60, Appendix A, Section 3.1, instructors could consider the use of AI as cheating under the following statement: "having ready access to and/or using aids or devices (including wireless communication devices) not expressly allowed by the instructor during an examination, test, quiz, or other evaluation."

Policy 60 is currently under review and will likely have specific guidance on artificial intelligence.

For more information on Academic Integrity and AI, visit Artificial Intelligence FAQs.

Speed of Information

The area of artificial intelligence is rapidly changing. While this guide will be periodically updated, new information may be available that is not currently reflected here.

Artificial intelligence, as defined by the chatbot ChatGPT, refers to "the development of computer systems that can perform tasks that typically require human intelligence, such as learning from experience, understanding natural language, and making decisions based on data" (OpenAI, 2023).

There are many different types of artificial intelligence, but the AI we think of most commonly today is generative AI (GAI). Generative AI learns from existing information in order to fulfill a prompt or request. For instance, if you ask a GAI tool like ChatGPT to write a birthday card for a family member, it will write a message based on the data it has been trained with while also learning from your prompt and any information you have provided.

Different Types of AI

Reactive machines follow specific predefined rules and inputs but do not learn or adapt over time. Some real-world examples are: chess programs like IBM's Deep Blue, home robots like the Roomba, traffic light control systems, simple chatbots, or Netflix's recommendations.

Limited memory machines use past and present experiences to make more effective decisions. They can learn from more data but can't retain data over long periods of time. Some real-world examples are: large language models like ChatGPT, self-driving cars, predictive text and autocorrect, virtual assistants like Siri or Alexa, and facial and voice recognition.

Theory of Mind AI and Self-Aware AI are two theoretical AI functionalities that have not yet been realized. Theory of Mind focuses on machines understanding thoughts and emotions. Self-Aware AI refers to machines that have a level of awareness about their own existence and capabilities.

Learn More

This short list of AI tools are mostly free but may have some limitations, including that they may:

• Only provide basic responses to common queries but may struggle with more complex interactions
• Provide lower quality answers
• Have limited integration capabilities
• Have limited or no dedicated support
• Limit the number of queries in a given time
• Have advertisements

Text Generators

Image Generators

Writing & Study Aid

Other Tools & Resources

This guide has been created by the Toronto Metropolitan University Library and is licensed under a Creative Commons Attribution International 4.0 License unless otherwise marked.