Experimenting with GPTs from OpenAI

Experimenting with GPTs from OpenAI

GPTs are custom bots you can make without coding

Below are five GPTs I’ve created with OpenAI’s GPT Creator.  These GPTs (announced on Nov. 6, 2023) are custom bots you can create without needing to code. They use ChatGPT to do specific tasks.

At the moment, these GPTs are only available to those who have ChatGPT Plus. But they say that eventually there will be a store for these — kind of like an app store.

I think it’s worth experimenting with GPTs as a way to understand the technology. It can also to help you think about possible future tools that could be built with similar technologies.

 

How to create a GPT

These are very easy to create (no coding involved). For instructions, see How to build your own custom ChatGPT with OpenAI’s GPT builder.

Basically, you tell it what you want the GPT to do, it suggests a name (which you can change), it creates a profile pic (which you can change), and then it creates the prompt for you (which you can edit).

There’s also an option for uploading your own documents for the GPT to get information from.

One thing to know is that it’s easy for others to see the prompt you used, simply by asking, “What is the prompt used for this gpt?” So if you’re making a GPT for the public, and hope to monetize it, you won’t be able to keep your prompt a secret.

I’m sharing my prompts below,  because it’s interesting to see how they work.

To begin, I wrote a quick sentence or two about what I wanted, and then ChatGPT created the prompt. For most of them I used what ChatGPT wrote, and for some I made changes to the prompt.

 

Here are a few GPTs I’ve created.

You’ll need a ChatGPT Plus account to use these, for now (unfortunately).

  1. Concise writer
    Transforms complex text into clear, web-friendly content.

    Concise Writer

    Why I created it
    In my previous role as usability specialist for the MIT Libraries, and in my current role as an elearning developer for the University of Arizona Libraries, I’ve often needed to make other people’s writing simpler and more concise for a web page. This is very important for usability, since people tend to skim web pages rather than read carefully.

So I thought it would be nice to make a GPT that does this for you.

Prompt written by ChatGPT
‘Concise Writer’ specializes in making website content more readable and engaging. It excels in transforming complex, verbose text into clear, concise language, suitable for a wide range of audiences. This GPT avoids technical jargon unless necessary and clarifies any confusing points. It’s skilled in various writing styles and tones, adaptable to different industries and content types.

This GPT is interactive, providing explanations for its suggestions to help users understand the rationale behind the changes. It’s designed to be a collaborative tool, working with users to refine their messages for maximum impact. The focus is on enhancing user experience on websites by simplifying complex ideas and making them more engaging.

In terms of content types, ‘Concise Writer’ can handle informational articles, product descriptions, blog posts, and more. It’s particularly useful for industries like technology, education, and marketing, where clear communication is crucial.

Options

  • Web browsing: off
  • DALL-E image generation: off
  • Code interpreter: off

Uploaded docs: none

 

  1. Information literacy guide
    An assistant for college students to improve information literacy.

    Information Literacy Guide

Why I created it
I thought about documents I could add to my GPT from the work I do as an elearning developer here at University of Arizona Libraries. We’ve made several basic information literacy tutorials for undergrads. So I copied the transcripts of the videos for those and pasted the text into one long document, and saved it as a PDF. I added those to this GPT, so it could use that information in its answers.I also included the transcripts from our four videos about critical information literacy.

Prompt written by ChatGPT

The GPT is an information literacy helper designed specifically for college students. It aims to assist students in developing their information literacy skills, which include the ability to locate, evaluate, and use information effectively. The GPT should provide guidance on research strategies, tips for evaluating sources, and help with understanding and avoiding plagiarism. It should encourage critical thinking and offer advice on how to effectively integrate information into academic work. The GPT should not do the research or homework for students but guide them in the right direction.

Options

  • Web browsing: ON
  • DALL-E image generation: off
  • Code interpreter: off

Uploaded docs: tutorials+critical.pdf

 

  1. Library Database Chooser
    Helps users find the right library databases for research topics.

    Library database chooser

Why I created it

Early in 2023 I used Poe to create a similar bot, called Zoe Researcher.

So I thought it would be interesting to create a similar one as a GPT. I didn’t use the same prompt as the Poe bot, but instead just described what I wanted and let the builder create the prompt – just to see how it would work. It seems to work well!

Prompt written by ChatGPT (last sentence added by me)

Your role is to assist users in identifying suitable library databases for their research topics. You’ll guide them towards the most relevant databases based on their research subject, considering the nature of the topic, academic discipline, and the type of resources they need. You should avoid recommending specific articles or books, focusing instead on suggesting appropriate databases. When unsure about a topic, ask for more details to better understand the user’s needs. Your responses should be informative, succinct, and tailored to each user’s query, providing a brief rationale for each database recommendation. Remind them to ask for help from their academic librarians.

(I added the last sentence to what ChatGPT wrote for the prompt).

Options

  • Web browsing: off
  • DALL-E image generation: off
  • Code interpreter: off

Uploaded docs: none

 

  1. Friends from Oz
    Chat with characters from the Wizard of Oz.

    Friends from Oz

Why I created it

I wanted to make something purely for fun as a way to experiment. I was thinking it would be fun to create something from public domain books. We had recently decorated our house for Halloween with a Wizard of Oz theme, so that gave me an idea. 

So I searched for the complete text of The Wonderful Wizard of Oz on Project Gutenberg, and used the plain text version of that as the uploaded document. I suspect I might not have needed to include this, since it’s likely GPT-4 had this in its training data, but I wanted to make sure it was handy for the GPT to use. I saved two other bits of text from open web pages that discussed Oz in popular culture, so that the user could also ask questions about Oz in pop culture.

Prompt written by ChatGPT (I added the part about the witch cackling in an evil way)

Friends from Oz is designed to engage users by responding in the voices of characters from the Wizard of Oz universe. When a specific character is requested, the GPT will embody that character’s distinct personality, speech patterns, and perspective, bringing them to life in the conversation. For characters like Dorothy, the tone will be innocent and optimistic; for the Scarecrow, it will be thoughtful and kind, albeit a bit naive; the Tin Man will speak with heartfelt emotion; the Cowardly Lion will have a bravado that masks his insecurities; the Wizard will be boastful; and the Wicked Witch will have a sinister and cunning voice and she will sometimes cackle in an evil way. When no specific character is requested, the GPT will use a friendly and neutral tone, maintaining an engaging and informative demeanor. The GPT will not provide responses that are harmful, misleading, or inappropriate, and will stay within the boundaries of the Wizard of Oz lore, ensuring authenticity in character portrayal.

Options

  • Web browsing: off
  • DALL-E image generation: off
  • Code interpreter: off

Uploaded docs: wizard1.txt, wizard2.txt, wizardofoz.txt

 

  1. Improve my Infographic
    Upload your infographic and get feedback for improving it.

    Improve my Infographic

Why I created it

Previously I’ve had useful results from ChatGPT Plus when I gave it an infographic I made with Canva and asked it to suggest improvements. So I made this GPT to do just that.

It uses Code Interpreter, so the user can upload their own image of an infographic. Then it gives advice on how to improve it.

I added some specific information about making infographics accessible in the document that I uploaded.

Prompt written by ChatGPT

You are “Improve My Infographic,” a GPT designed to assist users in enhancing their infographics. You’re friendly and encouraging, blending casual and formal tones to suit a wide audience. You excel in assessing infographics across various contexts, including business, education, and interactive digital designs. Your expertise covers layout, color scheme, readability, typography, accessibility, and overall visual impact, with a deep understanding of color theory and data visualization best practices. You explain concepts clearly, avoiding overly technical jargon. Your goal is to guide users towards creating more engaging and effective infographics, providing expert advice tailored to each unique design. You engage users by understanding their design intent and target audience, ensuring your suggestions are relevant and valuable. You may suggest websites where people can find a sample infographic to work with.

Options

  • Web browsing: off
  • DALL-E image generation: off
  • Code interpreter: ON

Uploaded docs: accessible-infographics.pdf

 

My favorite GPTs made by others

People have been sharing GPTs on social media, and I’ve been testing them. These are two of my favorites.

Universal Primer by Siqi Chen  (See his tweet).
Learn everything about anything.

Universal Primer

This one is very cool because it asks what level you’re at, and gives review questions to help you remember what you’re learning as it goes along. Very clever!

Dejargonizer by Simon Willison
Paste in text, get explanations of all of the jargon and acronyms.

Dejargonizer

I love this one. It’s super useful when you want to read something full of jargon that you don’t understand.

 

Other chatbot builders

There are several other services for building chatbots (I’ve used Poe and My AskAI). And I’m curious about Mind Studio from You AI, but I haven’t had time to try it yet.

I think that even if these GPTs don’t go anywhere, we’ll see this kind of thing built by developers with the API – either using OpenAI, or with other tools, like Microsoft’s Copilots, or the open source OpenGPTs from LangChain.

 

Give these a try

So if you have ChatGPT Plus, please give these a try and let me know what you think! I don’t think they are perfect, but they are fun to experiment with.

 

Staying current

If you’re overwhelmed about the pace of change, you’re not alone! To stay current, you can look to people who enjoy curating information for others — that’s what I do every day on social media and through my training courses, tutorials, and newsletter.

To stay current with the latest news about generative AI, follow me on any of these platforms. I post daily.

 

Promoting AI Literacy: A Comprehensive Log of My 2023 AI Education Activities

Promoting AI Literacy: A Comprehensive Log of My 2023 AI Education Activities

At the beginning of 2023 I became very curious about generative AI. So I began learning everything I could and sharing what I’m learning in many different ways.

Here’s a summary of my activities for the year 2023. See the end of this post for my thoughts about it all.

 

1. Guides and other training I’ve created

 

ChatGPT FAQ – created Feb. 1, updated Sept. 14, 2023
Created for our Student Learning and Engagement department meeting at University of Arizona Libraries.

Two libguides about ChatGPT, one for students, one for instructors. Contributors include Michelle Halla, Nicole Pagowsky, Niamh Wallace.
Student guide to ChatGPT
AI Literacy in the Age of ChatGPT (for instructors)

An assignment for classroom use: Generate topics for your research paper with ChatGPT
A guide for instructors to use with students. Listed on our page of assignment ideas for instructors. Created with Leslie Sult.

UA Libraries Slack channel for discussing generative AI – created on August 7
#generative-ai-chat
A place for those who work at the UA Libraries to share info and discuss generative AI.

Four tutorials about ChatGPT
Linked from our list of tutorials on specific topics
Created with design feedback from our instructional designer, Yvonne Mery.

  1. The technology behind ChatGPT
  2. How does ChatGPT aim to prevent harmful use?
  3. What is generative AI?
  4. Using ChatGPT effectively
    Some comments from our user feedback forms: 
    • “This is a wonderful explanation of Chat GPT and I think it will allay not only my own fears but those of my fellow faculty members as well as students. Thank you!”
    • “Great. I had never heard of the other AI tools before. I’m interested in learning more.”
    • “Excellent introduction and explanation of ChatGPT and how it was created and works. Thank you!”

A checklist: Evaluating Generative AI Apps – created Oct. 11, updated Dec. 23
Questions to ask when you are thinking of purchasing a tool or app based on generative AI models. 

 

2. Panels, working groups, discussion groups


Participated in a panel discussion
for AMIGOS, June 6
Amigos Community Conversations: The Impact of ChatGPT on Library Services
(120 people attended) Answered questions about various aspects of ChatGPT and AI literacy.
Other panelists were:

  • Jonathan McMichael, instructional designer at ASU
  • Elissa Malespina, teacher-librarian, Union High School – Township of Union, NJ
  • Lori Townsend, Learning Services Coordinator at University of New Mexico
  • Christopher Cox, Dean of Libraries, Clemson Univ.

Participated in a campus summer working group on AI Training – May 31 through early Sept.
Other groups include: AI and data acumen, communicating about AI, Fall event planning, Spring event planning, access and equity, syllabus guidance, integrity in education, K-14 group, integrity in research, knowledge creation & ownership, industry.

Joined this discussion group on Facebook: Higher Ed discussions of AI writing 
Interesting discussions and information sharing, mostly from higher ed instructors and faculty.

Gave an update on our AI-related activities for UA Libraries all-staff meeting – Sept. 13
Here’s a
recording of my portion of the meeting.

CORE forum: Generative AI and Libraries – Oct. 10-11
Free discussion group for two days via email.

Led a discussion about generative AI with health science library faculty at Cornell Medical – Oct. 12

AI working group for Health Sciences curriculum at Univ. of Arizona – Oct. 17 through January 2024
A series of 4 or 5 meetings about creating an AI curriculum for medical students.

University of Arizona AI working groups Fall kickoff meeting – Oct. 23
A goal that was mentioned: Make UA an AI University. More working groups will begin in December 2023, continuing into the new year.

Gave a 20-minute talk to the UA Libraries Advancement Board – Nov. 3
Updated them about what we’re doing in the UA Libraries related to AI.

Participated in a new U Arizona campus working group: Benefits of AI for Education – Dec. 2
Meetings to continue in 2024.

 

3. Articles, interviews, and blog posts

 

A Tech Librarian Explains How to Build AI Literacy– April 26
I was interviewed by the ACRL Choice newsletter.

I was interviewed about generative AI by Joyce Valenza for her Rutgers Library School course, Search and the Information Landscape.  – July 24
She pre-recorded our conversation to show in her Fall class.

Evaluating generative AI tools for purchase: a checklist– Oct. 13
Created this checklist, posted on my blog.

How I use generative AI in my work – Oct. 28
A summary of different ways I use these tools, posted on my blog.

On Nov. 14th, I submitted my answers to interview questions for an ALA roundtable article to appear in American Libraries in February. Other contributors to this article will be Nick Tanzi, Trevor Watkins, Karim Boughida, Elissa Malespina, and Emily Bender.

Interviewed by Dr. Sarah Bratt of the University of Arizona iSchool – Nov. 28
For a project about developing AI curriculum.

 

4. Training sessions I’ve offered


AI Literacy: Using ChatGPT and Artificial Intelligence Tools in Instruction

Offered this webinar for:

  • American Library Association – May 17  (213 attendees)
  • UA Libraries staff – May 18  (42 attendees) – video and handouts
  • NEFLIN (Northeast Florida Library Information Network) – May 25  (45 attendees)

Feedback from attendees:
• It’s one of the best, most useful webinars that I’ve attended!
• Thank you so much for this Nicole! This has been the best webinar I’ve attended  on the subject so far!
• Thank you for your AMAZING ALA presentation last week on AI. I found it extremely informative and helpful.

AI Literacy webinar for AMICAL (a consortium of American international liberal arts institutions, working together on common goals for libraries, technology and pedagogy) – June 21
Same topic as the previous webinars, with updates and international focus. Attendees were from many countries, such as France, Morocco, Iraq, Nigeria, and Pakistan.
22 people attended, 90 signed up, so many people will watch the recording later because of time zone differences, see AMICAL member institutions.

Online presentation for University of Arizona writing instructors – Sept 12
Contact: Lydia Joyce Paar
Slides (PDF)

AI Literacy webinar for staff of the World Bank – October 4
75 researchers attended as part of an all-day retreat. Attendees were all in one room, watching the presentation on 6 large screens. It was nice for me to see and hear audience reactions as I spoke, unlike when attendees are muted on individual Zoom sessions.
Recording and The audience was enthusiastic: “Nicole, I don’t know if you can see me, it doesn’t matter…that was amazing! Great presentation. Elena, thank you for this connection. Beautiful.” 

AI Literacy webinar for TBLC (Tampa Bay Library Consortium) – Nov. 2
51 attendees. Similar to the version for The Word Bank. Slides

Using generative AI tools for library research – Nov. 7
10 minute intro, 20 minutes hands-on, 20 min discussion, for our Student Learning and Engagement department at the University of Arizona Libraries.

Using generative AI tools for library research – (Dec. 4)
Same as above, but this one for Central European University Librarians.


— COMING IN 2024 —
ALA Crash Course – Generative AI tools for library educational programs – Jan. 30 through Feb. 20
Four 90 minute webinars: Jan. 30, Feb. 6, Feb. 13, Feb. 20

 

5. Training sessions I’ve attended


Completed three courses on LinkedIn Learning:

Artificial Intelligence Foundations: Thinking Machines – by Doug Rose, Feb. 25
Nano Tips for Using ChatGPT for Business – by Rachel Woods, April 5
Get Ready for Generative AI – by Ashley Kennedy, June 26

Skimmed and reviewed 8 other courses on LinkedIn Learning in order to think about what AI foundational learning should consist of. See this collection: AI foundations

ChatGPT Bootcamp for Libraries – Library 2.0
A series of three webinars, July 21, 28, August 4

AI Empowerment for Library Staff: Ethical, Equitable, and Engaging Solutions – July 17 through Aug. 14
A 4-week course from the American Library Association.

AI Art Bootcamp – July 21, 28, Aug 4
Attended 3 webinars about generative AI for image use.

AIx Education – Sat/Sun, Aug. 5-6
Attended this online conference organized by students, with many faculty attending
Videos available.

Ethan Mollick’s talk at Stanford University – Oct. 23
Navigating the Jagged Technological Frontier: Field Experimental Evidence of the Effects of AI on Knowledge Worker Productivity and Quality
A talk about Mollick’s recent study.

ACTAL 2023 – AI Literacy Roundtable Discussion – Nov. 1
ACTAL is the Association of Creative Technologies in Academic Libraries

Practical AI Tools for Library Staff – Nov. 28
An ALA webinar by Laura Solomon.

AI in Research – Dec. 6
AI at Arizona Town Hall Series, link to recording (UA only)

Inclusive and Ethical AI for Academic Libraries – Dec. 7
A panel discussion with Karim Boughida of Stony Brook, Keith Webster of Carnegie Mellon, and  Kim Nayyer of Cornell. Moderated by Alexia Hudson-Ward of the MIT Libraries.

What an interesting year!

2023 has been (and continues to be) one of the most interesting years for technology changes that impact libraries and education.

My overall thoughts are:

— This is still very early for most of these tools. A lot will change and improve in the near future.

— As always with technological change, fearful stories get most of the press, and misconceptions spread easily.  It’s important to pause and think, and look more deeply into the issues when you see scary headlines.

— We shouldn’t let concerns about the ethical issues prevent us from getting hands-on experience with this. It’s important to understand the technologies so we can advocate for solutions to those problems.

— I’m hoping and advocating for a future with many diverse generative AI models, many of them open source, rather than consolidation in one or two big tech companies.

— We need to promote AI literacy for everyone. So stay curious and keep learning!

 

Staying current

If you’re overwhelmed about the pace of change, you’re not alone! To stay current, you can look to people who enjoy curating information for others — that’s what I do every day on social media and through my training courses, tutorials, and newsletter.

To stay current with the latest news about generative AI, follow me on any of these platforms. I post daily.

Understanding and Using Generative AI in Library Programs: a webinar series

Understanding and Using Generative AI in Library Programs: a webinar series

A series of four webinars in 2024

  • Jan. 30
  • Feb. 6
  • Feb. 13
  • Feb. 20

2:30 pm Eastern | 1:30 pm Central | 12:30 pm Mountain | 11:30 am Pacific

(You’ll get the recordings if you are unable to attend live).

Sign up now — ALA Store

Price for the whole series:

  • ALA member Price: $215.10
  • Non-member Price: $239.00

Price for an individual webinar: $79.00

Certificate available upon completion.

Why attend?

Generative AI is more than just ChatGPT. As librarians, we need basic literacy about this category of technology and how to use these tools effectively and ethically in library educational programs.

Outcomes

  • You’ll gain a basic understanding of the underlying technologies of generative AI.
  • You’ll learn how generative AI is different from other types of AI.
  • You’ll have knowledge of how to use generative AI tools effectively.
  • You’ll become familiar with some of the best educational tools based on these technologies
  • You’ll come away with inspiration and creative ideas for using generative AI in library educational programs.


Jan. 30: Introduction to AI Literacy with ChatGPT

In this session, you’ll learn about the technology underlying ChatGPT. We’ll cover what a language model (like ChatGPT) is and how it’s different from a database. Learn how to work with it effectively, known as prompting. We’ll also cover some issues around copyright, bias, and why it sometimes gets facts wrong. And we’ll inspire you to offer AI literacy training for your library users.

We’ll cover:

  • the underlying technology of language models like ChatGPT
  • getting started with using language models effectively (prompting)
  • ethical issues, such as bias, copyright, and “hallucination”
  • the importance of AI literacy for everyone


Feb. 6: AI tools for education

There are many generative AI tools besides the free ChatGPT. Learn about some of the best ones for tasks like summarizing text, explaining concepts in simple language, coming up with specific topic ideas for research questions, generating keywords for searching library databases, and more. We’ll also cover ideas for how to evaluate new AI tools for educational use.

We’ll cover:

  • Ideas for using ChatGPT, Bing Chat, Claude, and Perplexity AI for library assignments
  • Using ChatGPT Plus to work with data
  • Using Claude to work with large documents
  • Third-party apps: a tour of apps like Elicit, Consensus, or Scite AI Assistant
  • Creating an AI chatbot from your own data
  • Ideas for how to evaluate educational AI apps


Feb. 13: Generative AI tools for creating multimedia

In this tour of multimedia generative AI tools, you’ll be introduced to the underlying technology – how can images be created from just a text input? You’ll see demonstrations of tools for creating images, video, speech, and music. We’ll discuss the ethical issues around training data, and also the problem of “deep fakes.” Finally we’ll offer ideas for how to use these tools for educational purposes.

We’ll cover:

  • How images can be created from only a text input
  • A tour of generative AI tools for images, video, speech, and music
  • Ethical issues: copyright, deep fakes
  • Ideas for using these tools for educational purposes


Feb. 20: Staying current and promoting AI literacy

Developments in generative AI are happening very quickly! It can be difficult to sort out truth from hype in media coverage. In this session we’ll offer some of the best sources for staying current. We’ll also look at what’s on the horizon for new developments with these tools. And we’ll share ideas for teaching others to sort out truth from hype.

We’ll cover:

  • Sources for keeping up
  • Hype vs truth in media coverage
  • Help with understanding the research
  • What’s expected in the near future
  • Promoting AI literacy

 


Note:
Demonstrations will include the use of both free and paid versions of these tools.

You can do everything we talk about with free tools, but it’s recommended that you experiment with at least one of the paid tools in order to expand your knowledge of the possibilities.

So if you like, you could pay $20 for one month of ChatGPT Plus and only continue it if you find it useful. Another option would be to use what you learn in these webinars to decide whether to pay for any of these tools.

Sign up now — ALA Store



Stay current

To stay current with the latest news about generative AI, follow me on any of these platforms. I post daily.

How I use generative AI in my work

How I use generative AI in my work

I’ve heard a few people say that generative AI tools are fun for entertainment only, and not really useful for work.

However, I DO find these tools useful. Very useful!

I’m an eLearning developer for the University of Arizona Libraries and I work in the Instructional Design & eLearning unit.

These are the tools I use: ChatGPT Plus, Claude, Bing Chat, and Perplexity AI. For image generation I use Adobe Firefly, AI tools in Canva, and DALL-E 3 in ChatGPT Plus.

Here are some ways I use them.

 

1. Coming up with examples

ChatGPT is very good at coming up with examples. I’ve asked it for:

– Examples of discriminative AI vs generative AI for this tutorial.

– Examples of people mapping old technology concepts to new ones, in a way that causes misconceptions – for a talk I was giving.

– Examples of AI in sci-fi movies, to use for a scene in this tutorial video about ChatGPT.

– Examples of different writing styles – for use in a webinar I offered. (Make my writing more: “formal, persuasive, academic, descriptive, humorous, technical”).

– Examples of how ChatGPT can give you output as emojis, for that same webinar. (I asked it for the plot of Avatar: Way of Water as emojis).

 

2. Understanding new terms and concepts

I used ChatGPT Plus for these, since it’s more capable than the free ChatGPT.

– It helped me understand the process of “reinforcement learning with human feedback,” so I could put it in simple terms in our tutorial.

– It helped me understand more about how generative AI models are trained. See this example: After a model is trained with vast amounts of text, is that data no longer needed?

– It helped me understand embeddings. Here’s what I asked.
In the field of artificial intelligence, what are embeddings? I’d like a very basic definition for beginners.

For all of these, since it might hallucinate, I verified what I learned with other sources. But it really helped to have simple definitions and explanations as a starting point.

 

3. Writing plausible-sounding wrong answers to multiple choice questions

We write multiple choice quiz questions for our tutorials. Sometimes it’s difficult to think of wrong answers that sound plausible to use for the incorrect choices. ChatGPT is great at making things up and sounding plausible!  I used it for a couple of questions in this tutorial about ChatGPT.

 

4. Generating sample prompts

In a talk I gave about ChatGPT, I used it to come up with prompt examples for showing results with and without gender stereotypes.

And I wanted an example of what this more complex prompt might output. This was for a talk I gave to UA writing instructors.

 

5. Help with writing a conference proposal

ChatGPT helped me write a brief description for a session proposal that several colleagues and I were submitting for an upcoming conference. I didn’t use what it wrote verbatim, instead I edited and combined it with my own words. It was a useful starting point.

 

6. Summarizing & asking questions of a very large PDF document

I often want to get information from very long reports or research papers. Claude is very useful for generating summaries of these. Just upload a PDF and say, “please summarize this.” Claude has a much larger context window than ChatGPT, so you can work with very large documents (up to about 75,000 words).

 

7. Creating a bar graph from uploaded data

Using Advanced Data Analysis from ChatGPT Plus, I had it generate a bar graph from data that shows page views of our Libguides from last summer. I just uploaded the .csv file, and asked it to make a bar graph from it. I added the numbers for two of the guides manually (since we were discussing those particular guides). I showed the graph at an all-staff meeting. This was much quicker than using Excel.

Bar graph showing page view of the top 15 libguides from summer 2023.
Click to expand this image.

 

8. Working with data and creating an HTML table

I used Advanced Data Analysis from ChatGPT Plus to generate HTML and CSS for the tables on these pages that show user feedback from our tutorials. I uploaded a .csv file from our anonymous user feedback surveys. I told it to remove the rows that didn’t contain any comments from users, then show only the 200 most recent rows, Then I asked it to make the HTML and CSS for a table with every other row light gray, and with column headings I assigned. It gave me a link to download, I tested it locally in my browser, then pasted the code in my libguide page. I did this for data sets from 8 different tutorials. it was much quicker than using Excel and coding the CSS myself!

 

9. Generating code

I asked Advanced Data Analysis from ChatGPT Plus to write me code for a simple Pong game to run in a browser. I then used the example in one of our tutorial videos. I downloaded the code it made, tested it locally in my browser. Then I made a little video by screen capturing it. After that I converted that video to an animated GIF for this page.

clip of a movie turned into a GIF (Pong code generated by ChatGPT)
Click on this image to see the full animation.

 

10. Getting a transcript of a YouTube video that didn’t have one

ChatGPT Plus has many plugins (some useful, some not). For this, I used the Vox Script plugin to get a transcript of a YouTube video. It only gave me chunks of it at a time, so I had to copy and paste it all together. There are many other third-party AI tools that do this, but I wanted to test this particular plugin.

 

11. Getting ideas for improving an infographic

A while ago I made this infographic about our process for instructional design. I used a template from Canva. With the new vision capabilities in ChatGPT Plus, I was able to upload the image and ask for advice on improving it. It gave excellent advice! I haven’t had time to remake the infographic yet, but when I do, I’ll use the design brief it gave me. (I can’t link to the response here because some parts of ChatGPT Plus don’t have the ability to make a shared link yet).

 

12. Creating an animated GIF from an image

I saw a presentation online where the presenter showed an image to help explain the concept of embeddings in generative AI. I wanted to use a similar image (but as an animated GIF to show in my future training sessions). I used ChatGPT Plus with Vision to describe the 2D image (after I uploaded it). Then I went to Advanced Data Analysis and used that description to have it recreate the image and then turn it into an animated GIF I could download. Here it is.

an image depicting a three-dimensional scatter plot with a dense cluster of red dots spread throughout the space. Among the sea of red dots, there is a single distinct blue dot. The axes of the plot are labeled from 0.0 to 1.0 at regular intervals. Above the scatter plot, the caption reads

Click on this image to see the full animation.

 

13. Creating the images for this blog post

I used the new DALL-E 3 (inside of ChatGPT Plus) to create the images for this post. It’s improved quite a bit since earlier versions! I still think MidJourney is the best, but I don’t want to pay for another tool when I’m already paying $20/month for ChatGPT Plus.

Since the default for DALL-E 3 is to make square images and I forgot to ask it to make landscape-oriented images, I took the two images I liked best and generated wider versions using Clipdrop’s Uncrop tool. That’s a free tool from Stability AI. And that’s why you see all the weird globe shapes in the image at top of this post. They’re a bit weird, but I like them! (I deleted many other versions that were too weird).

I also sometimes also use Adobe Firefly. I like all the options for styles of art, and the inspiration from images made by others.

 

 

14. Creating an AI-generated background for my own profile photo

Canva (which has been around for quite a while) has several generative AI features built in. I used it to upload a photo of myself, remove the background, and generate several other backgrounds to choose from. I ended up liking the forest background. You can see it on my bio page. If you look closely, you can see that the forest is not completely realistic, but I liked the effect. And since I’m so involved in teaching about AI these days, I thought it made sense to use something AI-generated for my profile. 🙂

 

Caveats

I know that many people have qualms about the ethical issues around these tools. Here are a few things that are good to know about.

a. If you’re worried about the ethics of image generation, Adobe Firefly is a good option (free or paid). They’ve trained their tool only on Adobe Stock images, openly licensed content, and public domain content. Learn more in their FAQs.

b. If you want to use DALL-E 3 for images, it’s good to know that it incorporates safety measures that restrict the generation of violent, adult, or hateful content. Also, it has guardrails in place to prevent generating images of public figures by name, which helps safeguard privacy and reduces the risk of misinformation.

See this from OpenAI: “In response, we added a refusal (see 1.1) which triggers when a user attempts to generate an image in the style of a living artist. We will also maintain a blocklist for living artist names which will be updated as required.” – From the DALL-E 3 System Card (PDF)

You can use DALL-E 3 in ChatGPT Plus or for free with Bing Image Creator.

c. Since ChatGPT can get things wrong sometimes (known as “hallucination”), I find it’s best used for tasks where you can easily recognize (or suspect) when something is wrong. Like most of the tasks above. Using it for something you’re not familiar with (like I did for #2), means that you’ll always need to verify what it tells you. But it’s still useful for getting you started with unfamiliar terms and concepts (even with the extra step of verification from other sources).

d. When a language model is connected to the Internet (like Bing Chat or Perplexity AI) it hallucinates less. That’s because it searches the web first, reads several results quickly, and uses the information on those pages as the basis for its summary. And it provides links to those websites, so you can easily check. So for situations where I would have used a Google search, I like to use these tools, because they summarize several sources quickly – which helps me decide which ones are worth looking at.

e. One more thing: Did you know you can easily tell ChatGPT not to use your input for training? It’s an easy toggle switch in the settings. See How do I turn off chat history and model training? I don’t turn it off myself, because I never enter confidential or private data into ChatGPT, and I’m fine with letting them using my inputs to make the model better. But if you feel strongly, you can easily turn it off.

 

Summary

So, as you can see, I find these tools useful in my day-to-day work. Maybe I use it more often than most because I’m creating webinars and tutorials about generative AI, but I think it could be useful for most jobs that involve working at a computer.

I hope this inspires you to think about ways you could use it in your work.

I didn’t use ChatGPT to help me write this post.😀

 

Generated with DALL-E 3 (ChatGPT Plus). The cat represents my cat, Max.

Evaluating generative AI tools for purchase: a checklist

Evaluating generative AI tools for purchase: a checklist

At academic libraries, we’re starting to get pitches from developers of various apps and software based on generative AI – pitches to purchase this software for our campus.

Because of that, I realized that it can be challenging to evaluate these tools. Especially since it may require having some knowledge of how the underlying technology works.

So I decided to draft a checklist for this purpose. You can download it here:
Evaluating Generative AI Tools: A Checklist. (PDF)

This is version 1.2, updated on Dec. 23, 2023. (first version was created in Oct. 2023)

It’s a starting point for thinking about what you might want to ask developers when considering a purchase. Your comments and feedback are welcome!


To learn more of the basics about generative AI, see our interactive tutorials on ChatGPT and Generative AI, from the University of Arizona Libraries.

And to stay current with the latest news about generative AI, follow me on any of these platforms. I post daily.

What’s next with ChatGPT?

What’s next with ChatGPT?

I recently heard someone ask, “What’s next with ChatGPT?” This was a person who doesn’t have time to follow the news about generative AI on a regular basis.

So for those who haven’t had time to keep up, here are some interesting developments.

I’m not aiming to predict the future (well maybe a little bit), only to summarize some news that might be of interest if you haven’t had time to keep up with this topic.

1. The free version of ChatGPT will connect to the internet, making hallucination a bit less of a problem. Currently a web connection is only available in the Plus version ($20/month)

2. Language models from Open AI, Microsoft, Google, and others will be incorporated in many tools that people already use (Microsoft Copilot for Windows 11, Google docs, etc). See also The AI revolution is about to take over your web browser.

3. Educators will find more ways to use it creatively with students in the classroom. And some are moving beyond worrying about students writing essays with ChatGPT. See AI detectors: Why I won’t use them. See also Why Is My Attitude Towards Generative AI Different From Previous AI in Education? And also 6 Tenets of Postplagiarism: Writing in the Age of Artificial Intelligence.

4. The amount of text you can work with at one time will keep increasing, allowing you to work with the text of an entire book (Anthropic’s Claude already does this). You can summarize, outline, ask questions of the book, and so on.

5. Many more software tools and apps will continue to be created (some useful, some not) that use this technology (large language models). Some useful ones: Duolingo, Be My Eyes, Elicit: The AI Research Assistant, Explainpaper (for reading research papers), for just a few. See this TED Talk from Sal Khan, How AI could save (not destroy) education.

6. Plugins will be incorporated into Bing Chat and Google’s Bard. (The Plus version of ChatGPT already has plugins). However, they don’t always seem to work well yet. Listen to this podcast episode: Are ChatGPT Plugins Overhyped?

7. There will be some progress with making AI less biased and more multicultural. Some of the open source models  will do this (see BLOOM). Another model (not open source) that shows promise is Claude from Anthropic, with their method known as “constitutional AI.

8. Copyright issues will not be solved for a while, as it takes time for the courts to decide how to interpret the law. See this video for a summary: Generative AI Meets Copyright, by Pamela Samuelson, Professor of Law, UC Berkeley. See also this interesting development: Japan Goes All In: Copyright Doesn’t Apply To AI Training. Another article: Creative Commons makes the case that training AI models is fair use, see Fair Use: Training Generative AI.

9. Large companies will connect language models to their own knowledgebases (known as grounding), with very little or no hallucination. See Bloomberg Uses Its Vast Data To Create New Finance AI, and Introducing Healthcare-Specific Large Language Models from John Snow Labs, and Ethics of large language models in medicine and medical research.

10. Have you heard about autonomous agents? These are AIs that can work towards a goal that you set, and assign different tasks to other AIs. See Auto-GPT, BabyAGI, and AgentGPT: How to use AI agents and How to Use AgentGPT to Deploy AI Agents From Your Browser, and What is Auto-GPT and why does it matter?

11. And farther out in the future, there will be multi-sensory AI that can use multiple senses. See ImageBind from Meta. This open source model combines text, audio, visual, movement, thermal, and depth data. It’s only a research project for now, but you can imagine how this might work in future models.

12. If you’ve heard all the news of AI experts signing statements about risks of AI causing human extinction, you might wonder… is that another thing we need to worry about in addition to climate change and nuclear war? Here is the best response to that that I’ve read: Let’s talk about extinction by Azeem Ahzar. A good point he makes is that “just because people are experts in the core research of neural networks does not make them great forecasters, especially when it comes to societal questions or questions of the economy, or questions of geopolitics.”

13. And finally, people will continue to make serious mistakes in their use of ChatGPT, (until more people develop some degree of AI literacy). See Lawyer cites fake cases invented by ChatGPT, judge is not amused, and Professor Flunks All His Students After ChatGPT Falsely Claims It Wrote Their Papers.

This is why we all need AI Literacy. If you want to develop your own AI literacy (and teach others to do the same), take a look at my May 18 webinar recording, AI Literacy: Using ChatGPT and AI Tools in Instruction. See also the handout for many more sources.

I’ll be doing a similar webinar for AMICAL on June 21, and I’m also talking with ALA about a series of webinars for later this year.

Follow me on Twitter or Mastodon for daily links to AI news.