1 of 78

AI Strategies for AI Strategies for Student Support Staff

Lance Eaton

This work is licensed by Lance Eaton under a Creative Commons Attribution-ShareAlike 4.0 International License.

Image Source: DALL-E 3 from ChatGPT4

2 of 78

What is generative AI?

Wrong answers only…

Image Source: ChatGPT

3 of 78

“Generative AI is a magical creature that lives inside your computer and makes all your dreams come true. It spends its days knitting digital sweaters and baking cookies for other programs. In its spare time, it enjoys composing symphonies for squirrels and painting abstract portraits of invisible friends. The primary function of generative AI is to serve as a professional online fortune-teller, predicting whether you'll have spaghetti or tacos for dinner based on the alignment of your browser history.”

Image Source: ChatGPT

Now, when I asked ChatGPT for wrong answers only.

““Generative AI is a magical creature that lives inside your computer and makes all your dreams come true. It spends its days knitting digital sweaters and baking cookies for other programs. In its spare time, it enjoys composing symphonies for squirrels and painting abstract portraits of invisible friends. The primary function of generative AI is to serve as a professional online fortune-teller, predicting whether you'll have spaghetti or tacos for dinner based on the alignment of your browser history.””

Frankly, I think we did better.

Thank you for that. I like to start us with this activity because well, a little levity can help us just release different things we’re holding onto and let us be ready to see things different.

4 of 78

ChatGPT

Claude

Bing/CoPilot

Gemini/Bard

Barnabus

Perplexity

WriteSonic

What tools do you know about?

Image Source: ChatGPT

Here are some of the currently most popular AI tools. Let’s see which ones folks have heard of.

But by show of hands, which ones of you have heard of…

ChatGPT

Claude

Bing

Gemini/Bard

Barnabus

WriteSonic

We do see a variety of experience in the room. Ok, does anyone want to describe any of these?

And no cheating by going to ChatGPT to get an answer!

You can learn more about these on the annotated slide deck.

These are some of the most popular tools but generative AI is soooo ubiquitous in just this last year–its in our google and microsoft products, it’s on our phones, its in our grammar checkers!

It’s so hard to make sense of it all and keep up that I just made up Barnabus and everyone just assumed it was actually an AI tool they just hadn’t heard of yet.

And it’s easy to dismiss this as another tech-hype cycle. I think in many ways it is. But I also think the AI Hype feels different from prior cycles.

5 of 78

Image Source: iFake Text Message

The Textbox As a Portal to Generative AI

Generative AI is different from the AI we’ve been using for the last decade.

We all have stories of how Siri or Alexa has misheard us. These didn’t feel as impressive or easy to use right out of the box. But generative AI–AI that can generate text, audio, image, or video through the use of everyday language–that feels different because of its ease of use.

We’re several years since the launch of the Metaverse and can ANYONE tell me what the hell the Metaverse is in 50 words or less? Like–tell it to me as if I were a 5 year old?

The lift to figure out what the Metaverse is, how to access it, how to create in or with it, and why it would be better than other things–that’s a lot of figure out.

But the generative AI comes in the form of a chatbox–a chatbox many of you have already been using in this very zoom meeting…something that’s been around for decades. We all know: see a textbox on the computer, enter text.

Generative AI took something terribly complex and nuanced and made it usable in a textbox. That’s a devilishly easy invitation.

6 of 78

Category #1:

Efficiency

7 of 78

Efficiency Examples

Task minimizer
Accessibility
Adapting content
Generating lists (dates, process, etc)
Prioritizing tasks

8 of 78

Prompt: Task Minimizing

Prompt #1

Create a verified list of all the Tuesdays between August 18, and December 21, .

List them like this example: Tuesday, August 19, .

Prompt #2

Provide a verified list of any US holidays that happen between those two dates.

List them like this example: [Holiday Name] - Monday, September 2,

9 of 78

Prompt#1: Response

Tuesday, August 20,

Tuesday, August 27,

Tuesday, September 3,

Tuesday, September 10,

Tuesday, September 17,

Tuesday, September 24,

Tuesday, October 1,

Tuesday, October 8,

Tuesday, October 15,

Tuesday, October 22,

Tuesday, October 29,

Tuesday, November 5,

Tuesday, November 12,

Tuesday, November 19,

Tuesday, November 26,

Tuesday, December 3,

Tuesday, December 10,

Tuesday, December 17,

10 of 78

Prompt #2

Labor Day - Monday, September 2,

Columbus Day - Monday, October 14,

Halloween - Thursday, October 31,

Veterans Day - Monday, November 11,

Thanksgiving Day - Thursday, November 28,

11 of 78

True or False

Generative artificial intelligence may be becoming sentient.

Image Source: ChatGPT

12 of 78

FALSE

Generative artificial intelligence may be becoming sentient.

False. It is not a thinking being–no matter what all those sci-fi novels, movies, comics, TV shows, and Silicon Valley Tech-Bros tells us. We have a long history of telling ourselves that our objects are real–going back millenia.

It is a complex technology with lots of possibilities, but sentience is not one of them in the immediate future. What we hear and see in the media is often sound bites and misrepresentations of what generative AI is.

This is important to hold in our minds. This tendency to anthropomorphize or mythologize our technology as having life is common throughout history going back thousands of years. In more modern times, this has become known as the ELIZA effect after experiments in the 1960s with a computer program that people became attached to when it used language in ways that felt real to the users.

We want to remember this effect as we use it.

13 of 78

True or False

Generative artificial intelligence is similar to human thinking when it arrives at its answer.

Image Source: ChatGPT

14 of 78

FALSE

Generative artificial intelligence is similar to human thinking when it arrives at its answer.

15 of 78

How does AI arrive at its answers?

Large Language Model (LLM): The large data set needed to train the generative AI. For text generative AI, it has typically included large parts of the Internet

Token: The smallest unit of data needed to create the mathematical relationships within the dataset. 4-5 characters in length.

Generative (as in “generative AI): The process of calculating answers based on probabilities and statistics relationships between the prompt and the data set.

When working with text, a generative AI will rely on its “large language model”--that is a massive data set of text and the statistical models that calculate the relationships within that data set.

That large language model was created by an AI tool mathematically analyzing the relationships across that massive text. It doesn’t analyze the words themselves but rather strings of characters. Those strings of characters are referred to as “tokens” and are short, something around 4-5 character long.

The AI calculates the probability of relationships of those tokens across its data set and also with any prompt you give it. So when the AI gets a prompt by a user, it’s mathematically examining the relationship among tokens and then searching its large language model for similar relationships of tokens to help generate the most probable answer. It strings these tokens into words and sentences that mathematically make sense.

It’s not thinking, it’s mathing. It has no awareness or internal understanding of the relationships of the words its using any more than a calculator knows that its also spelling out a word with the number 376616.

Now you’re all wondering what word does that number spell? It spells giggle. Which, I’m pretty sure some of y’all just did!

16 of 78

Category #2:

Brainstorming

17 of 78

Brainstorming Examples

Generating different activities for learning
Planning
Concept mapping
Self-Assessment
Dialogue partner

18 of 78

Prompt: Brainstorming Part 1

Create a role-play scenario for an urban university sophomore who needs to discuss changing their major from Psychology to Communication with their academic advisor. Include potential advisor questions about the student's motivations, concerns about credits already earned, and long-term career goals. Provide guidance on how the student can articulate their decision and address potential challenges.

19 of 78

Prompt Response

Advisor: Good afternoon! I understand you wanted to discuss your major. How can I help you today?

Student: Hello, thank you for meeting with me. I've been thinking a lot about my academic path, and I'm considering changing my major from Psychology to Communication.

Advisor: I see. That's a significant change. Can you tell me more about what's motivating this decision?

Guidance for the student: Be prepared to articulate your reasons clearly. Focus on your interests, strengths, and how you've come to this decision.

20 of 78

Prompt Response

Student: Of course. While I've enjoyed my psychology courses, I've found myself more drawn to the communication aspects of human behavior. I'm particularly interested in how people convey ideas and the impact of media on society. I believe a Communication major would align better with my interests and career goals.

Advisor: Thank you for sharing that. It's good that you've given this some thought. I have a few questions to help us navigate this potential change:

How do you see your psychology credits fitting into a Communication major?
Have you considered how this change might affect your graduation timeline?
What specific career paths are you envisioning with a Communication degree?
Have you spoken with any faculty or students in the Communication department?

Guidance for the student: Address each question thoughtfully. Be honest about any uncertainties and show that you've researched your options

21 of 78

Prompt Response

Student: Those are great questions:

I've looked into it, and I believe some of my psychology credits could count as electives or fulfill general education requirements. I'm hoping we can review this together.�
I understand there might be some impact on my timeline, but I'm prepared to take summer classes if necessary to stay on track.�
I'm interested in public relations, digital marketing, or possibly journalism. I feel these align well with my strengths in writing and analyzing social trends.�
Yes, I've spoken with Professor Johnson from the Communication department and sat in on a Media Studies class. I found it fascinating and it reinforced my interest in the field.

22 of 78

It works well when we flip the script

Image Source: DALL-E 3 with ChatGPT4

PRO TIP

23 of 78

True or False

Generative artificial intelligence can lie.

Image Source: ChatGPT

24 of 78

FALSE

Generative artificial intelligence can lie.

25 of 78

True or False

Generative artificial intelligence will take all our jobs.

Image Source: ChatGPT

26 of 78

FALSE

Generative artificial intelligence will take all our jobs.

As Public Enemy said, “Don’t believe the hype.”

It will disrupt and augment work and part of the challenge for all of us will be to stay abreast of it to understand its impact on our areas of work. Any knowledge worker is going to have to consider and contend with how these tools can amplify their work and not replace it.

It will help us do a lot of things we didn’t do before and that can create new opportunities.

And yes, there are jobs that will be lost and there will be new jobs that emerge. The history of the world is filled with technologies that would supposedly ruin all employment and yes, there’s still plenty of work to be had.

Many of us are at educational institutions and we offer degrees and programming in many kinds of work that did not exist when we were born or have terminated programs that no longer have relevance. Technology continues to drive a lot of those professional changes.

27 of 78

Category #3:

Content Generation

28 of 78

Content Generation Examples

Outlines
Strategic projects
Assignment guidelines
Rubrics
Course Guides
Quizzing
“Student” Examples

29 of 78

We need 2 buckets

Source: DALLE-3 with ChatGPT 4

PRO TIP

Now, it’s create that it can produce content but that’s also a big concern.

There’s one definitive fact we know about generative AI for the foreseeable future. It calculates the probability of responses but it does not know the answers. Therefore, in order to use it well–really well–the kind of well that can mean a matter of life and death, useful and asinine is users need two buckets.

The first bucket is a working understanding of how generative AI works; its limitations, its capabilities, its biases, and how it fits into the larger cultural landscape of late-stage capitalism. That’s essential to knowing what it is and isn’t, can and can’t do–but more importantly, sifting through the promises of a lot tech-bros in Silicon valley.

The second bucket is a deep knowledge of the subject matter you’re working with it on. It just doesn’t know things; it only calculates responses and that can only get one so far. When it comes to disciplinary knowledge, it still falls drastically short. Yes, it can produce facts but it often can’t produce knowledge.

30 of 78

True or False

Generative artificial intelligence only has information since 2022 when it was released.

Image Source: ChatGPT

31 of 78

FALSE

Generative artificial intelligence only has information since 2022 when it was released.

32 of 78

True or False

Generative AI detectors can determine if something was created with generative AI.

Image Source: ChatGPT

33 of 78

FALSE

Generative AI detectors can determine if something was created with generative AI.

False. False. False. Generative AI detectors only calculate a probability. They do not actually prove anything. This means if you accuse a student of using generative AI–you do not have a good means to prove it. ��It’s different from plagiarism detectors which can point to something that existed. AI plagiarism detectors don’t work that way. And importantly, they have a significant false-positive rate. This means they will inevitably falsely accuse students and students have little means of proving innocence because we aren’t able to see the backend of such tools.

Of those that are likely to be accused–it’s people who are English-language learners that are more likely to be falsely accused. That’s because the AI plagiarism detectors are looking for patterns–and, of course, that’s how we learn and reproduce language is through patterns.

34 of 78

Avoidance is not a strategy

Image Source: DALL-E 3 with ChatGPT4

PRO TIP

35 of 78

Category #4:

Data & Info Processing

36 of 78

Data & Info Processing Examples

Analyzing research
Analyzing qualitative & quantitative data
Reviewing & summarizing documents
Comparing documents

37 of 78

Prompt: Qualitative Analysis

Examine the attached article on online tutoring, and explain how its findings are relevant to tutoring at a university. Also, provide direct and applicable actions that staff or tutors should do based upon this research. Be sure to explain why these actions are important and the intended result.

38 of 78

Prompt Response

This research article on online peer tutoring in higher education provides several key findings that are highly relevant to tutoring at universities:

Online peer tutoring positively influences academic performance across different disciplines. This suggests universities should implement online peer tutoring programs broadly, not just in select subjects.
Regular engagement in online peer tutoring is associated with higher student retention rates. Universities should encourage consistent participation in peer tutoring to improve student persistence.
Online peer tutoring enhances students' sense of community and increases engagement in coursework. This indicates peer tutoring can help combat feelings of isolation in online/remote learning environments.
Online peer tutoring improves collaborative learning skills among college students. This develops important teamwork abilities that benefit students academically and professionally.
Online peer tutoring is linked to higher graduation rates. This demonstrates its potential as an intervention to support degree completion.

39 of 78

Prompt Response

Based on these findings, here are some recommended actions for university staff and tutors:

Implement a structured, campus-wide online peer tutoring program: Why: To provide academic support across all disciplines and improve overall student performance. Intended result: Improved academic outcomes and student success rates across departments.
Establish regular tutoring schedules and encourage consistent attendance: Why: Regular engagement is linked to higher retention rates. Intended result: Increased student persistence and reduced dropout rates.
Incorporate collaborative activities and group discussions in tutoring sessions: Why: To foster a sense of community and develop collaborative learning skills. Intended result: Enhanced student engagement and improved teamwork abilities.
Provide training for peer tutors on effective online communication and engagement strategies: Why: To maximize the community-building and engagement benefits of online tutoring. Intended result: More effective tutoring sessions that keep students motivated and connected.

40 of 78

Always ask what it is missing in its answers

Image Source: DALL-E 3 with ChatGPT4

PRO TIP

41 of 78

Category #5:

Communication

42 of 78

Communication Examples

Editing text
Revising text
Review others’ communications for tone or clarity
Improve your own communications for conciseness, tone, etc

43 of 78

Prompt: Communication Part 1

Compose a script for a student with an invisible disability to use when discussing their accommodation needs with a skeptical professor. Include strategies for explaining the impact of the disability on academic performance, addressing potential pushback, and assertively reiterating legal rights.�

44 of 78

Prompt Response

Professor [Name], thank you for meeting with me. I wanted to discuss some accommodation needs related to my disability. I have [name of disability], which is not visible but significantly impacts my academic performance."

Explaining the Impact

Student: "My disability affects me in the following ways:

[Specific impact on coursework or class participation]
[Another specific impact]
[A third specific impact if applicable]

These challenges make it difficult for me to [specific task] without accommodations.

Keep an eye out for these things. Generally, this should not be shared with an AI.

45 of 78

Prompt Response

Addressing Potential Pushback

Professor: "I'm not sure I understand. You seem fine in class. How do I know this isn't just an excuse?"

Student: "I understand your concern. Invisible disabilities can be confusing. While I may appear fine, I'm dealing with internal struggles that aren't obvious. I've been formally diagnosed, and the Disability Services Office has verified my condition and approved these accommodations."

46 of 78

Prompt Response

Proposing Accommodations

Student: "To help me succeed in your class, I'm requesting the following accommodations:

[Specific accommodation]
[Another specific accommodation]
[A third accommodation if applicable]

These will help level the playing field and allow me to demonstrate my true abilities in your course."

47 of 78

Prompt Response

Addressing Further Skepticism

Professor: "I'm worried this might give you an unfair advantage over other students."

Student: "I appreciate your concern for fairness. However, these accommodations don't give me an advantage – they remove barriers that my disability creates. They allow me to access the course material and demonstrate my knowledge on an equal footing with my peers."

Asserting Legal Rights

Student: "I want to emphasize that these accommodations are protected under the [relevant law, e.g., Americans with Disabilities Act or Section 504 of the Rehabilitation Act]. The university is legally required to provide reasonable accommodations for students with documented disabilities."

48 of 78

Only I We need to figure this out

Bing Image Generator with prompt: “students and faculty standing together staring at a giant computer monitor that has a robot face on it looking back at them”

PRO TIP

Now, this is the hard one for academia--society as a whole, but particularly academia.

None of us can really figure this out on our own; WE need to figure it out together.

Yes, any one of us can decide what we're going to do in our courses, implement that, and go about our business.

But this is a significant change--it's a change on par with the pandemic and the rise of the digital classroom.

We need our collective wisdom--the wisdom that can bring different knowledge and understanding about these tools, the ways to use them, the things that work, won't work, or might work.

AI is here and is going to continue to change and no single person can keep up with it all. I sure as heck can.

Everything that I'm able to do is because I do it in communities of practice with colleagues, friends, professional organizations, and other folks online.

49 of 78

One Final Prompt

The prompt to start ALL AI chats:

“Improve the following prompt to maximize the creativity and analytical capabilities of a large language model:

“[prompt]”

Ask it first to improve the prompt

FINAL

PRO TIP

50 of 78

Questions

This work is licensed by Lance Eaton under a Creative Commons Attribution-ShareAlike 4.0 International License.

Image Source: DALL-E 3 from ChatGPT4

51 of 78

10 minute break!

52 of 78

Testing out Prompts

(15 minutes)

53 of 78

Step 1

Go to Resource document�

Select the Resources related to your area

54 of 78

Step 2

Open up another tab.

�Go to any of these:

https://claude.ai/chat/

(account creation req.)

https://chat.openai.com/

https://www.perplexity.ai/

55 of 78

Step 2 (Advanced)

Open up a third tab.

�Select a second AI tool:

https://claude.ai/chat/

(account creation req.)

https://chat.openai.com/

https://www.perplexity.ai/

56 of 78

Creating An Account

Benefits

Still free
More choices to use
Keep track of chatlogs
Can share chat-logs

Drawbacks

More data sharing

57 of 78

Step 3

�Review the prompts on your team’s resource guide.

Try at least 3-5 prompts

58 of 78

Step 4

�Based upon results, try to create a few prompts of your own to try out

59 of 78

Prompt Library Reminders

Categories are not perfect�
Prompts for staff, student workers, & students�
Play with the prompts as needed

[Personalized brackets]�
Imperfect answers=retry�
Example-seeking: tell it to repeat

Ask follow up questions�
Challenge outputs�
Turn it into an interviewer when unsure

60 of 78

Office Discussions

(10 minutes)

61 of 78

Step 1: Document�

Designate a note-taker
Use a collaborative document (Google Docs, Word, etc)
Capture key pieces of the conversation

62 of 78

Step 2: Discuss�

Share observations
What worked?
What aha moments did you have?
What do they want to return to for their specific work?

63 of 78

Divisional Share out

(10 minutes)

64 of 78

Share-Out�

Each group share 1 insight or interesting usage you discovered

65 of 78

Office Collaboration

(15 minutes)

66 of 78

Step 1: Document�

Designate a note-taker
Use a collaborative document (Google Docs, Word, etc)
Capture key pieces of the conversation

67 of 78

Step 2: Identify an Objective with AI�

Based upon group usage, determine what is a goal that you want to achieve for using AI that you would all agree to.

68 of 78

A Word about Objectives

Consider using SMART goals to guide the final objective

Approaches

Build something with AI
Build a practice around AI
Build guidance for AI with students

Examples

Refined Prompt Library
Specific object to build (guide, templates, strategy, plan)
Guidance on team usage

69 of 78

Step 3: Email the full goal to Lance and (Division Leader)�

Send: [email]

Subject: AI Workshop Goals

Include: Your Office & Goal

70 of 78

Step 4: Finish Early?�

Get started on the next phase (see resource document)

71 of 78

Project Work

(15 minutes)

72 of 78

Step 1: Group up within your areas

No more than 3 people in a group (e.g. groups fo 2-3 people)

73 of 78

Step 2: Build It

In that smaller group, work towards building out your objective

74 of 78

Office Debrief

(10 minutes)

75 of 78

Share results of groups

Briefly demo where you go to and what you discovered

76 of 78

Division Discussion

(20 minutes)

77 of 78

Share out

2 minutes to share

What your goal was
What you build
What you noticed in the different examples

78 of 78

Thank you!

This work is licensed by Lance Eaton under a Creative Commons Attribution-ShareAlike 4.0 International License.

Image Source: DALL-E 3 from ChatGPT4