What Is Generative AI?

The answer to that question may vary from source to source and the "lens" at which it is being viewed or used. Let's look at what the University of Nebraska – Lincoln Center for Transformative Teaching provides.

Generative Artificial Intelligence is a program that can create "new" content by using and referencing existing material. These are programs that "listen" to songs, "read" articles, and "see" art and then create a new piece of material based on the query posed to it. It is important to note that generative artificial intelligence does not generate new ideas or work. Instead, it uses information derived from existing works (often many) to find the average or most common pathway to create the content asked of it.

Source: https://teaching.unl.edu/resources/strategies-techniques/teaching-technology/ai-gen-ai/Links to an external site.

Please browse this UNL site further, as you may discover a few thoughts to explore for your own teaching and learning. If you have previously viewed this resource, I encourage you to re-scan through it, as the content has changed a bit over the past few months.

GPT = Generative Pre-Trained Transformer:

"Generative AI uses a mathematical model to identify and encode data patterns and then predict which word should come next."

"This is also why detection works, because this model creates a distinct statistical signature that looks very different from what humans tend to create in their writing. We can detect that difference."

Eric Wang, Turnitin's Vice President of AI

Generative Predictive Text:

It's a software-coded mathematical model to identify text, encode patterns, and then predict which words should come next.

As educators, can we not account for this mathematical model to create learning experiences and assignments to demonstrate competency and attainment of learning outcomes via various "other" means?

Let's set student up for their success! Create assignments that empower creating "new" knowledge vs regurgitating our past.

What is the educational challenge in Higher Education?

We're amidst a shift in action as a reaction to emerging software tools.

from

"How do I know if my student is submitting plagiarized work?"

to

"How do I know this assignment was or was not written by Ai?"

Early systems detected plagiarism by clearly identifying the specific source of work copied from. With Generative AI, there are multiple sources for every word selection, and those infinite sources will constantly change. There is no objective evidence for its results. Educators are discovering new and exciting ways to leverage AI to engage students, provide deeper learning, save time, and streamline tasks.

Work is underway by many public and private partnerships to provide tools to both assign and preserve original authorship metadata as well as tools to later "inspect" content and dig deeper to support faculty in those hard-to-have conversations. One such initiative can be found at https://contentauthenticity.org/Links to an external site. and an example at https://contentcredentials.org/verifyLinks to an external site.. Users can drag and drop content to explore metadata behind resources and content.

This tool takes me back to lessons I taught regarding the validity of internet content and its original sources. Leveraging the Way Back Machine at https://archive.org/web/,Links to an external site. you can browse many resources and "travel back in time" to explore how those sources have originated and evolved. For educators, a helpful feature enables one to capture a web page as it currently appears so it can be used as a trusted citation in the future. I cannot count how many times I've linked to a web resource only to discover later my students cannot access it as I intended.

Using Generative AI

Numerous tools have been developed to generate various types of data, such as Scribe, Jasper, ChatGPT, Dall-E2, Autodesk's Generative Design, Wordtune, Notion, GitHub Copilot, VEED, and Speechify. Adobe FireflyLinks to an external site., OpenAI's DALL-E 2Links to an external site., and MidjourneyLinks to an external site. can produce images based on user text prompts. Soundful.comLinks to an external site., which is AI-generated audio, is very intriguing as well. Many other technologies leveraging all forms of AI, including generative AI video, are emerging every day.

Educational organizations are working diligently to provide educators with resources to help make sense of it all and provide resources to leverage the tools for application in teaching and learning successfully. TeachAI.orgLinks to an external site. and it’s AI Guidance for Schools ToolkitLinks to an external site. is one such resource created collaboratively by numerous non-profit educational organizations.

I attended an AI-focused webinar hosted by Instructure, the parent company of Canvas. Justin Hodgson from the University of Indiana leveraged Adobe Express to create a web-based resource in lieu of a slide deck for his talk on Generative AI, Chat GPT & Syllabi: An Ethics of Practice. With his permission to share with you, please take a few minutes to browse this resource at https://express.adobe.com/page/DjrK4zSIgCey1/Links to an external site.

It is often the case where emerging technologies are introduced into the business world and then trickle down into innovative educational uses. Form follows function, and of course, follow the money, and you will find the treasure. Amazon is providing an AI-powered image-generation tool to help its advertisers create better ads. Take a quick peek at https://www.aboutamazon.com/news/innovation-at-amazon/amazon-ads-ai-powered-image-generatorLinks to an external site. Now I can remember spending hours creating just the perfect image for my teaching and learning materials. This tool from Amazon would have saved me a lot of time. To think of this tool being used in education. How would you use it to "up your game" and create a more engaging educational experience for your students?

What do you need to know to get started?
Simply put, a desire to try new things. Over the years, as technologies and internet services changed, we have reached a point where a search engine was just the tip of the iceberg. We are in an age of emerging reasoning engines that have exploded. Reasoning engines are software programs or logic systems that use algorithms to draw conclusions, make inferences, summarize information, and help solve problems based on available data.

Next we will explore prompt engineering, a digital literacy skill we will need to know and become fluent with.

Prompt Engineering

What is it?

Watch "Harness the power of prompt engineeringLinks to an external site." (4m 2s) from LinkedIn Learning. This is part of a larger learning path called Career Essentials in Generative AI by Microsoft and LinkedIn. Feel free to explore the whole path, but I’d like you to skip ahead into Section 2 to focus specifically on prompt engineering portion.

*Note: you will need to create a free LinkedIn account if you do not already have one.

If you choose not to create a LinkedIn account or have technical difficulties, please read the video transcript that I have exported and provided below.

Transcript: Harness the power of prompt engineering.

One of the most important skills to develop when working with reasoning engines is in the area of prompt engineering. There's truly an art to the way you craft your prompts when you query reasoning engines. - Prompt engineering is the act of asking questions of one of these foundation models. So the first time you go in and you enter in a prompt, you can consider yourself a prompt engineer. But what accelerates that capability and your ability to do a better job is actually telling the model, not just asking it a question, but then also giving it examples of the question and the answer you would like to see. We often call this one shot or few shot learning and what it means is how much instruction are you giving when you're asking it a question. So the prompt itself is really quite simple. It's just the question you ask. But the instructions that you provide to guide the answer, that is where true prompt engineering starts to kick in. - There is this incredible skill set that you have to develop to be able to harness these tools, and it really, really is relying on the development of appropriate prompts. You will get out only as good of a prompt that you put in. So it's almost like the great painters of the past were those who could wield the paintbrush. I think the greatest artist maybe in the future are those who can wield the prompt. Part of that tinkering in knowing, okay, well this is the image it gave to me from this prompt. I really want to move it in this way. How do I manipulate the wording of the prompt to get it to give me that? And that is such an amazing skill set that people are developing. - Now, let me tell you, this just takes practice. When I first started to enter prompts, I asked questions like I would ask a search engine and I didn't get great results. It wasn't until I learned how to hone my language that I started getting the results I was after. And as I continued honing my prompt language, I started getting results that I didn't even know were possible. Take a look at the difference between this prompt, write a social media post about International Women's Day and this one, write a cross section of social media posts for Facebook, LinkedIn, and Twitter for International Women's Day. My business is a boutique video editing company and I want to celebrate female film editors. Keep the tone celebratory, casual, fun, and optimistic. And then, how about this prompt? An image of a red car, and this one, an image of a red car in a busy downtown area approaching a traffic jam. The car should be the only color in the image. The rest should be black and white. In each of these examples, additional details regarding content, context, tone, and aesthetic, give me better results. Again, as you learn how your prompts impact the output you create, it's important to analyze the results for various factors, not just related to the content itself, but also to various quality standards. And if the results aren't up to your expectations, continue to hone the prompt. - Look at the results and evaluate are these results accurate? Are they discriminatory in any way? Are they only presenting one political viewpoint? And then I think it's continuing to iterate. The power of these systems is continuing to iterate on those prompts. It's not just one prompt and you're done. It's start from a prompt, see where it gets you and then keep regenerating and refining the prompt to get at a more accurate or better result. - [Instructor] Here are some resources for you to start learning the art of prompt engineering. Don't be shy, just start prompting, analyzing, and tweaking and I promise you'll improve the way you query over time to generate some awesome results.

Let's give prompt engineering a try!

While numerous AI tools exist, feel free to leverage any or all. I'll give you a couple to choose from to get you started.

Disclaimer: I have reached out to UNMC Information Security professionals, and while a formal statement regarding the educational use of AI is in the works, it is highly recommended that we:

- Explore at your own risk.
- Act within institutional acceptable use policies (University of Nebraska Executive Memorandum No. 16Links to an external site.
- Think professionally in all cases, and do NOT enter any sensitive information or personally protected health information.
- Use @unmc.edu to authenticate into work provided software and web services
- Use personal email addresses to create free accounts while exploring non-UNMC IT approved software and web services

Microsoft Copilot

- Visit https://copilot.microsoft.com
- To get started, you will need to login.
  - You are permitted to use your UNMC ID and password
- Type into the text box and press return/enter on your keyboard.
- What do you see, what do you notice?

ChatGPT:

- Visit https://chat.openai.com/Links to an external site.
- To get started, you will need to create an account
  - USE a personal email address to create this account
  - DO NOT enter any UNMC propriety or confidential information, as these tools can use that data in future training models.
- Leverage the free version
- Type into the chat-style text box and press return/enter on your keyboard
- What do you see, what do you notice?

Claude

- Visit https://claude.ai
- To get started you will need to create an account.
  - USE a personal email address to create this account.
  - DO NOT enter any UNMC propriety or confidential information, as these tools can use that data in future training models.
- Type into the text box and press return/enter on your keyboard.
- What do you see, what do you notice?

NotionAi:

- Visit https://www.notion.so/Links to an external site.
- To get started, you will need to create an account
  - USE a personal email address to create this account
  - DO NOT enter any UNMC propriety or confidential information, as these tools can use that data in future training models
- Leverage the free version
- Follow the on-screen prompts
  - For users of Microsoft OneNote, this will look/act very familiar
  - Press the spacebar and type into the text box
  - What do you see, what do you notice?
  - Explore the left-hand navigation options and the additional features button in the upper-right

What do you see? What do you notice? Compare and contrast these tools. Which one do you prefer over the other?

If you would like to explore Generative AI further in a fun, self-paced, and guided learning experience, check out Adobe’s Education Exchange professional learning course at https://edex.adobe.com/professional-learning/self-paced-course/generative-ai-in-the-classroom-with-adobe-express/overviewLinks to an external site.. (approximately 30 minutes)

How could you adapt what you discover in this experience to your current teaching and learning?

Take a Deeper Dive! A List of AI Newsletters: