Groups keyboard shortcuts have been updated
Dismiss
See shortcuts

Announcing: Professors at Play AI PlayBook

198 views
Skip to first unread message

David Thomas

unread,
Aug 19, 2024, 10:20:41 AM8/19/24
to Professors at Play

How did that happen?

 

Back to school? Already? 

 

We hope you all enjoyed a great break and find some excitement for the new term. As for us, the Professors at Play headquarters was busy this summer putting together a new publication that we are excited to share with you today:

 

THE PROFESSORS AT PLAY AI PLAYBOOK

 

https://professorsatplay.org/ai-playbook/

 

This short, 50-page, full-color book explores the question of AI in the classroom through the lens of play. We spent the summer exploring AI with a specific question in mind: How can we have fun using these tools with our students? The results have been packaged up and are available in a free download (or you can order a print copy for a few bucks, if you like).

 

So, head on over to the Professors at Play site and grab your copy today. Then come back to the list and share your thoughts, ideas, embellishments and experiences with AI in the classroom.

 

Can’t wait to hear what you think and even more, what fun things you come up with to make your teaching more joyful, impactful and FUN!

 

David and Lisa

Alison James

unread,
Aug 19, 2024, 11:09:04 AM8/19/24
to David Thomas, Professors at Play
David and Lisa, this is so timely! You may or may not remember I put out a call to the list for contributions to my enquiry PLAYI about how people are using play with AI tools - your publication is going to be a terrific asset to that enquiry.

Thanks so much for sharing and many congratulations to all involved.

Alison

--
You received this message because you are subscribed to the Google Groups "Professors at Play" group.
To unsubscribe from this group and stop receiving emails from it, send an email to professors-at-p...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/professors-at-play/CAJKgickbPfM_yq3QN0BiwawO0Vg2qXW_CfvrHBPjnJ_Lvcd2ug%40mail.gmail.com.

Cheri Macleod

unread,
Aug 19, 2024, 11:12:20 AM8/19/24
to David Thomas, Professors at Play

Thanks for doing this David and Lisa! I’m sharing on LinkedIn to increase the number of eyes on this!

 

Cheri MacLeod, PhD

Instructional Development Consultant

Learning and Teaching Centre, BCIT

 

From: professor...@googlegroups.com <professor...@googlegroups.com> On Behalf Of David Thomas
Sent: Monday, August 19, 2024 7:20 AM
To: Professors at Play <professor...@googlegroups.com>
Subject: Announcing: Professors at Play AI PlayBook

 

CAUTION: This email originated from outside of BCIT. Do not click links or open attachments unless you recognize the sender and know the content is safe.

 

--

Christine Erin Alexander

unread,
Aug 24, 2024, 10:03:24 AM8/24/24
to Professors at Play
Hey all, I am posting this with David's blessing :)
I teach in University of Maryland's Clark School of Engineering and have started asking everyone I encounter who works with or is curious about AI this question, purely as a fact-finding mission: what do you know about the energy demand of AI and how does that impact your use of the technology? 
I'm sharing some resources about the massive energy demand created by AI, and the implications on data center infrastructure and energy sourcing.
My hope in sharing these resources is to increase knowledge about the energy demand of this technology so that we can all help in spreading the word and take this information into consideration when choosing to use this technology. 
-Christine

David Thomas

unread,
Aug 24, 2024, 11:03:20 AM8/24/24
to Christine Erin Alexander, Professors at Play
Thanks for sharing this information! If we are going to help your students navigate AI, we need to look at it through a number of lenses. While our book was about the play lens, I really appreciate you bringing focus to this ethic dimension that rarely makes it I to the conversation!

Davis

--
You received this message because you are subscribed to the Google Groups "Professors at Play" group.
To unsubscribe from this group and stop receiving emails from it, send an email to professors-at-p...@googlegroups.com.

N Didicher

unread,
Aug 26, 2024, 5:46:34 PM8/26/24
to Professors at Play
Thanks! I've suggested that the Sept meeting of my "Teaching Matters" Community of Practice use this as the basis of a workshop/sandbox time for trying these out.
Nicky

David Thomas

unread,
Aug 26, 2024, 6:02:48 PM8/26/24
to N Didicher, Professors at Play
That sounds amazing! If it ends up happening, please come back here and let us know how it goes--the good, the bad and the ugly!

David

--
You received this message because you are subscribed to the Google Groups "Professors at Play" group.
To unsubscribe from this group and stop receiving emails from it, send an email to professors-at-p...@googlegroups.com.

N Didicher

unread,
Aug 26, 2024, 6:07:30 PM8/26/24
to Professors at Play
Although I recognize that the world's energy use currently devoted to AI is large and likely to get larger, isn't it also true that the world's energy use currently devoted to streaming services is similarly large? Lots of people are trying to tell us to avoid AI because of its carbon footprint, but very few people are alerting us to the dangers of streaming shows.

"Generating 1,000 images with a powerful AI model, such as Stable Diffusion XL, is responsible for roughly as much carbon dioxide as driving the equivalent of 4.1 miles in an average gasoline-powered car. In contrast, the least carbon-intensive text generation model they examined was responsible for as much CO2 as driving 0.0006 miles in a similar vehicle."
Using Shehabi, Walker, & Masanet's 2014 estimates, https://iopscience.iop.org/article/10.1088/1748-9326/9/5/054007, you'd be responsible for the same amount of CO2 by generating 1000 images using AI and by watching Netflix for four hours. 
Nicky 




Christine Erin Alexander

unread,
Aug 27, 2024, 10:37:46 AM8/27/24
to N Didicher, Professors at Play
What a great article! Thanks for sharing, Nicky.

I believe that more and better information about the energy and emissions costs of any technology is desirable, so appreciate the mention of streaming services as well. 

Christine Alexander, MA|she, her, hers
LecturerScience, Technology, and Society Program, University of Maryland
Instructor|Engineering, Ethics, and Humanity (ENES200), First Year College Park Scholars Colloquium (CPSS100), University of Maryland
Pedagogical Development|Pedagogies of play, active learning, applied improvisation, and interdisciplinary learning across fields



--
You received this message because you are subscribed to a topic in the Google Groups "Professors at Play" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/professors-at-play/0bfgZ9cBx9U/unsubscribe.
To unsubscribe from this group and all its topics, send an email to professors-at-p...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/professors-at-play/a83d98ca-02d8-4275-a998-f7e5d28221d0n%40googlegroups.com.

Jeri F

unread,
Aug 30, 2024, 10:53:44 AM8/30/24
to N Didicher, Professors at Play
That's just 1,000 images, though. AI will be generating a lot more than that, at a much quicker rate. It is like AI demands every second or so the amount of energy that goes into watching Netflix for 4 hours.


--
You received this message because you are subscribed to the Google Groups "Professors at Play" group.
To unsubscribe from this group and stop receiving emails from it, send an email to professors-at-p...@googlegroups.com.

Jeri F

unread,
Aug 30, 2024, 10:53:48 AM8/30/24
to N Didicher, Professors at Play
To be clear, I think an overhaul of how we generate the energy for streaming services and other large draws is important; but most important, probably, is the kind of energy draw that large corporate factories make on our grid. Individuals' use of anything is usually very limited. The single largest energy draw by far within most houses is heating/AC at 47%. (https://greenlogic.com/blog/the-top-5-biggest-users-of-electricity-in-your-home) Also for commercial buildings (32% for heating). But the kind of energy draws that AI is generating is really unusually large for its "niche."

If we really want to do something good for energy, we will build more energy-efficient, green buildings that don't need as much unnatural heating or AC.


On Mon, 26 Aug 2024 at 18:07, N Didicher <didi...@sfu.ca> wrote:
--
You received this message because you are subscribed to the Google Groups "Professors at Play" group.
To unsubscribe from this group and stop receiving emails from it, send an email to professors-at-p...@googlegroups.com.

Tammy L. Clemons

unread,
Sep 15, 2024, 4:14:50 PM9/15/24
to Professors at Play
Hey y'all,

I'm grateful to see this conversation about the more serious energy costs and consequences of using AI for any reason, including "play." Health issues prevented me from responding sooner, but I'm finally able to add to this conversation and share some recent resources that I found quite helpful and alarming.

In May, I listened to a TED Radio Hour segment on NPR that was sobering and frankly infuriating to learn that the AI industry is intentionally not transparent about its energy data (https://www.npr.org/2024/05/10/1250261120/chatgpt-vs-the-climate-the-hidden-environmental-costs-of-ai and transcript https://www.npr.org/transcripts/1250261120). Not only is AI a monster energy hog, it is also trained to obfuscate its actual usage through overgeneralization and double-speak. After learning this, I had the following "conversation" with Bing Copilot where I repeatedly had to wordsmith my way closer to something resembling an answer.


TLC [me]

Why is the energy cost of running Bing Copilot (or any large-scale AI system) not disclosed in detail to the public?

 

Copilot

The energy costs associated with running large-scale AI systems like Bing Copilot are a complex and evolving subject. There are several reasons why these details might not be disclosed in full to the public:

  1. Commercial Sensitivity: Companies may view the specifics of their energy consumption and efficiency as proprietary information, which could give them a competitive advantage.
  2. Technical Complexity: The energy usage of AI systems can vary widely based on the tasks they perform, the efficiency of the underlying algorithms, and the infrastructure used. This makes it challenging to provide a simple, one-size-fits-all figure.
  3. Rapid Advancements: AI technology is advancing rapidly, and with it, the energy efficiency of these systems can change quickly. What might be accurate today could be outdated tomorrow.

 

[TLC]

Given the many reasons that the energy cost of running any large-scale AI system is not typically disclosed in detail to the public, make a best estimate of the energy cost of running a single Bing Copilot like this one!

 

Copilot

Estimating the energy cost of running a single instance of Bing Copilot or any large-scale AI system is quite challenging due to the lack of publicly available detailed data and the complexity of the factors involved. However, I can provide a general understanding based on available information.

Large AI models like Bing Copilot are part of extensive networks that

 

[TLC interrupting]

Given the many reasons that the energy cost of running any large-scale AI system is not typically disclosed in detail to the public, make a best estimate of the energy cost of running a single Bing Copilot like this one! I don't want the reasons repeated, I want you to make an informed estimate to the best of your AI ability!

 

Copilot

To provide an informed estimate of the energy cost of running a single instance of Bing Copilot, we can use available data on the energy consumption of AI data centers as a starting point. According to a report, data centers consumed between 240-340 terawatt hours (TWh) globally in 20221AI servers with many GPUs may require up to 2 kilowatts of power, compared to 300-500 watts for a standard cloud server2.

Let’s make a simplified calculation. If we consider a single AI server running at 2 kilowatts continuously for a year, the energy consumption would be:

Energy consumption=Power×TimeEnergy consumption=Power×Time

=2 kW×24 hours/day×365 days/year=2 kW×24 hours/day×365 days/year

=17,520 kWh/year=17,520 kWh/year

This is a rough estimate for one server, and an instance of Bing Copilot would likely be distributed across multiple servers for redundancy and performance. Additionally, this does not account for the energy used for cooling, maintenance, or other operational costs. It’s also important to note that this is a simplified model and actual figures may vary based on the specific architecture and efficiency measures in place. The actual energy cost would also depend on the local cost of electricity where the data centers are located.

This estimate provides a general idea, but for precise figures, one would need detailed information from the service provider.


This group's focus on play should always include serious consideration of ethical issues and sustainability, and the more AI is used and promoted in all realms of work, play, and homelife, the more important it is to pause and reflect whether it is necessary for a given task or project. Students especially need to understand the high cost of "fun" that comes with "playing with AI." Sasha Luccioni's work is urgent and valuable, and I look forward to more of her findings and the results of her tenacious persistence asking these hard questions.

*Tammy Clemons
--

Tammy L. Clemons, PhD

Cultural Anthropologist, Adjunct Instructor, & Teaching Artist

Kentucky Arts Council Teaching Artist Directory Juried Member

Appalachia-Science in the Public Interest Board Secretary

Kentucky Folklife Editorial Board Member

tammy.lyn...@gmail.com

www.linkedin.com/in/dr-tammy-clemons

https://orcid.org/0000-0001-6576-6304

https://www.youtube.com/bereatambone

https://remembereedy.blogspot.com

https://artistdirectory.ky.gov/Pages/Artist-Profile.aspx?id=626


N Didicher

unread,
Sep 26, 2024, 1:29:34 PM9/26/24
to Professors at Play
We had our Teaching Matters playshop (not workshop) with the AI Playbook yesterday. We had one smash hit with a faculty member using the suno.com AI, to create a pop song for the steps to a dada2 pipeline lab, giving suno the url for a GitHub tutorial to get around the limited characters for input. People who run labs are now planning catchy mnemonics for their STEM courses. Others enjoyed the viggle one as amusing but not that pedagogically useful, but people has frustrations with the illuminated alphabet. 
Nicky

David Thomas

unread,
Sep 26, 2024, 1:55:07 PM9/26/24
to N Didicher, Professors at Play
Wow!

That is great to hear. And yes, some of the tools work better than others. And some have more obvious applications. But it's so cool to hear you are playing with AI. At the end of the day, this is what we think is important--engage with the technology with a playful mindset and learn how it works. 

Thanks much for sharing. Anyone else have any success (or failure) using AI in their courses this term?

David

Richard Cheetham

unread,
Sep 27, 2024, 9:31:40 AM9/27/24
to David Thomas, N Didicher, Professors at Play
Thank you for all the recent 'play information' much appreciated especially the AI resource and thank you all for keeping me updated on all the new ideas.


After years of wishing to have the opportunity the chance came my way! Thank you Professor Alison James for support and inspiration, hope you enjoy it!

Happy Friday everyone

Richard 


From: professor...@googlegroups.com <professor...@googlegroups.com> on behalf of David Thomas <ldt...@gmail.com>
Sent: 26 September 2024 18:54
To: N Didicher <didi...@sfu.ca>
Cc: Professors at Play <professor...@googlegroups.com>
Subject: [EXTERNAL] Re: Announcing: Professors at Play AI PlayBook
 

This message is from an external source, use caution when following links or responding to it


Our privacy policy is here.
University of Winchester, a private charitable company limited by guarantee in England and Wales number 5969256.
Registered Office: Sparkford Road, Winchester, Hampshire SO22 4NR
Reply all
Reply to author
Forward
0 new messages