GENERATIVE ARTIFICIAL INTELLIGENCE (GENAI) AT BYU

Frequently Asked Questions

• Toggle Item

  What is GenAI?

  Generative AI, or GenAI for short, is a subset of artificial intelligence that focuses on the generation of content when given a prompt. This could come in the form of image generation, text generation, and even video and audio generation.

• Toggle Item

  How does GenAI work?

  Generative AI, or GenAI, is a form of AI that absorbs vast amounts of data to create original content. This content could range from text to music or even images. It operates on complex machine-learning algorithms which are trained on extensive datasets.

  GenAI is presented with numerous content examples during training, helping it identify patterns and rules. Once trained, it can generate content that follows these identified patterns and rules.

  In simple terms, GenAI is an advanced mimic. Given enough examples, it learns to imitate the content convincingly, resulting in unique outputs similar to the original. Whether it's a new article, a music piece, or a graphic, GenAI helps create something unique and inventive.

• Toggle Item

  Where can I find training and other useful materials for learning how to use AI?

  You can find training materials on our GenAI training page.

• Toggle Item

  Can I use GenAI tools for academic assignments or research?

  Some academic departments have their own set of rules. You can find general information and guiding principles here, but otherwise, contact your department advisement for additional guidance.

• Toggle Item

  How do I get started using GenAI?

  There are many types of GenAI that are easily accessible. Find some here.

• Toggle Item

  What are prompts and what is prompt engineering?

  Prompt Engineering is a fascinating aspect of Generative AI. It's essentially the process of designing, fine-tuning, and optimizing the prompts or initial inputs that we provide to an AI system. The goal is to get the AI to generate the most accurate and useful output possible. Think of it like asking a question to a very knowledgeable friend, the way you phrase the question can greatly influence the type and quality of the answer you get. That's the essence of prompt engineering: it's about crafting the perfect 'questions' to elicit the best possible 'answers' from AI.

• Toggle Item

  Are there limitations to GenAI?

  There are many limitations to GenAI, and GenAI services should never be trusted as correct without human validation. Some of the limitations of GenAI include biased answers and completely incorrect answers (often referred to as "hallucinations"). Users should always fact check output from AI services to determine accuracy.

• Toggle Item

  What are some strengths and weaknesses of GenAI?

  GenAI is very good at certain things. For example, GenAI is very good at helping brainstorm ideas, checking for grammar errors in written text, and translating text into different languages. GenAI services also do a great job at role playing scenarios, which gives users valuable ideas and counter ideas. With that being said, there are also things that GenAI struggles doing. Weaknesses of GenAI include biased answers due to biased training data, as well as the possibility of low quality or incorrect answers. Understanding the strengths and weaknesses of GenAI can help users utilize it more efficiently.

• Toggle Item

  What are the ethical considerations of GenAI?

  There are many ethical considerations when using GenAI. Nonpublic Institutional Data should never be entered into a GenAI service. This is due to the fact that many services store user inputs for future training of the specific GenAI model. When GenAI has been used in any capacity, the generated content should be cited as AI-generated and the user should never attempt to take credit for AI-generated work. Lastly, output from GenAI models should always be verified. GenAI services are known to produce false or biased information and because of that, all output from GenAI services should be fact-checked before it is used in any capacity.

• Toggle Item

  Is Microsoft Copilot HIPAA and FERPA compliant?

  While BYU has a agreement with Microsoft for the use of Copilot, HIPAA and FERPA related data should not be entered into Microsoft Copilot. However, deidentification methods can be used to alter the data so that it is safe to use within Copilot, while still reaping the benefits of Copilot's capabilities. Training guides on deidentification data will be released on this site soon.