Introduction

The advent of Artificial Intelligence (AI) and large language models (LLMs) have ignited a lot of discussion in the Higher Education space. While some view AI as potentially disruptive and harmful (Educause, 2023; evidenced by initial bans in some school districts), others believe that it can transform teaching and learning for both educators and students (Educause, 2023). Bowen and Watson (2024) argued that the impact of AI may be more profound than past technological inventions like the internet. The invention of the internet, World Wide Web, changed our relationship with knowledge by increasing access to materials, thereby making knowledge more abundant. However, “AI is going to change our relationship with thinking” – “challenging ideas about creativity and originality” (Bowen & Watson, 2024, p.2).

What is Artificial Intelligence (AI) and Generative Artificial Intelligence (GenAI)?

Artificial intelligence (AI) is the ability of machines or applications to perform tasks that typically require human intelligence. Some common AI uses include voice assistants like Siri, Alexa, smart home devices, and facial recognition. Specific examples in education include Grammarly, Semantic Scholar, and Gradescope.  

Generative Artificial Intelligence (GenAI) refers to a subset of AI. GenAI, is a natural language model that can generate human-like responses using models trained on vast amounts of data. These responses could be text, image, graphic, video, or music.

Introduction to AI for Teachers and Students by Ethan Mollick and Lilach Mollick Video (10:16 mins)

How does GenAI work?

Large Language Models (LLMs) are trained on massive amounts of text data from sources like books, articles, websites, and social media posts. They learn patterns and structures from this data, allowing them to generate new content by predicting the most likely words or sequences. 

When you ask GenAI a question, you’re essentially giving it an instruction. This instruction could be something like: “Here’s a fragment of text (I like my bagel with…). Tell me how this fragment might go on. What words are likely to come next?” The GenAI then uses its understanding of language, which it has learned from the vast amount of data it has been trained on, to generate a continuation of the text.

It’s important to note that every output from GenAI is generated from the texts in its database. It doesn’t “know” anything in the way humans do, but rather, it uses statistical patterns to generate its responses. This means that while GenAI can produce remarkably human-like text, it doesn’t actually understand the content in the same way a human would. It’s a tool for generating content. Hence, users of GenAI (instructors and students) must review and double-check AI responses for accuracy, authenticity, and bias.

What is Prompt Engineering for GenAI?

Prompt engineering in GenAI involves finding effective ways to ask a language model (GenAI tools) to give you the information or response you want. A prompt is an input or instruction given to a language model to elicit a specific output. It is important to craft effective prompts to get a good response following the golden rule “garbage in, garbage out”. An effective prompt must include the appropriate phrases, words, and sentence structure to generate the best response. 

Bowen and Watson (2024) proposed features to consider when writing effective prompts. These included: 

• Task: What exactly do you want AI to do? – Create, brainstorm, predict, draw, generate, condense, summarize, translate, or assemble.
• Format: What is the specific output format? – Syllabi, lesson plan, essay, press release, opinion piece, dialogue, spreadsheet, table, or PDF.
• Voice: What style of language is desired? – In the style of a renowned scientist, like an editor, enthusiastic style, or empathetic style.
• Context: What further context or examples can you provide? 

Learn more about prompting A guide to prompting AI (for what it is worth) and view a video The Art of Prompting Engineering Simplified (10:08 mins)