Category Archives: AI

Transfigurations on Musings

Cosmic whale for cosmic writing

Hello, my friends. If you, like me, have ever gazed into the cosmos of thought and marveled at the boundless intersections of science, technology, and human understanding, then you’re in for a journey. Today, we embark on a thoughtful exploration inspired by the writings of Jennifer L. Adams—a thinker deeply entrenched in the realm of higher education, where technology and learning converge like celestial bodies in orbit. Her central question is both provocative and profound: Is artificial intelligence truly that different from how our own minds work?

Such a question beckons us to consider the intricate dance of memory, intelligence, and pattern recognition, and to marvel at their manifestations both natural and artificial. Adams begins her inquiry not in a laboratory or lecture hall, but in a bathtub—a setting both humble and evocative, echoing Archimedes himself. She watches whirlpools form and dissipate, contemplating the microscopic life swirling in these temporary eddies. Her curiosity takes her to a surprising discovery: slime mold.

Ah, slime mold—a single-celled organism seemingly so simple, yet capable of navigating mazes and anticipating environmental changes. Imagine: a creature with no brain, no nervous system, no neurons, yet it remembers. Its memory, Adams suggests, may be chemical, a fundamental organization of matter with purpose. It is here, in this primal intelligence, that we are invited to see echoes of artificial intelligence.

Adams draws a parallel to large language models (LLMs), like GPT-4. These models, too, operate without consciousness, yet they predict patterns and generate responses so human-like that they often blur the line between machine and mind. Consider this: when tasked with responding to a zoo worker’s query, the AI adapts, contextualizes, and personalizes its response. It mirrors the dynamic complexity of thought, much as the slime mold mirrors memory.

But Adams doesn’t stop at algorithms. She speculates on the broader implications of intelligence—animal, artificial, and human. She recounts the intricate songs of whales, passed down through generations, a kind of aquatic epic encoded in soundwaves. Could their communication represent an organic language model, evolved naturally and independently of human cognition? What might these songs tell us about their history, their emotions, their view of the universe?

This thought invites an even deeper question: if intelligence emerges in myriad forms—from the chemical traces of slime molds to the silicon networks of AI—what truly defines intelligence? Is it memory? Pattern recognition? Adaptability? Or something ineffable, like the capacity for wonder or the ability to ask questions about existence itself?

Adams provocatively ties this inquiry back to the classroom, to the very essence of learning. Imagine a world where AI personalizes education for every learner, a virtual tutor attuned to the unique pathways of each student’s mind. Yet here, she invokes a cautionary principle: the Prime Directive from Star Trek, a reminder that with great power comes great responsibility. How do we harness AI to amplify human potential without losing what makes learning an inherently human endeavor?

The bathtub becomes a metaphor for our role in this vast experiment. As Adams muses about pulling the plug, ending the microcosmic swirl of life, we are reminded of the fragility of discovery, the delicacy of choice. How we engage with AI, how we integrate it into education, and how we define its role in our society will shape not only our future but our very understanding of intelligence itself.

So, as we stare into the starry vastness of possibility, let us ponder: What if AI is not merely a tool, but a mirror? A mirror reflecting our own creativity, our capacity for connection, our endless curiosity? And in that reflection, perhaps we might better understand ourselves—not as isolated beings, but as part of a vast and intricate cosmos, forever learning, forever exploring.

Stay curious, my friends. The universe awaits.

Researcher? Apply to utilize Empire AI

Dear Stony Brook Faculty and Researchers:

Earlier this spring, New York State and six academic partners, including the State University of New York (SUNY), committed to create Empire AI*, a GPU supercluster for AI and high-performance computing academic research. Stony Brook faculty member Dr. Robert Harrison is serving as the interim director of Empire AI and can be a useful resource to faculty interested in exploring this resource. An early version of Empire AI’s research computing instrument, called “Alpha,” is coming online by early November 2024. The hardware specifications for Alpha are outlined at the end of this email message. While the ultimate Empire AI instrument will far outstrip Alpha’s specifications, our research communities can now start to use this shared resource. Alpha, should enable Stony Brook to do research we otherwise could not with existing resources.

Empire AI Consortium, Inc. welcomes SUNY  faculty and researchers to request time to run a research project using Alpha.
  • HOW: Please submit brief work order requests (WORs) to Empire AI via this secure online form. This short form asks for a paragraph description of your research project and five questions about your resource needs.
  • WHEN: You may submit your work order request starting October 3 through October 31, 2024. When you submit WORs within this time window has no bearing on when and whether the project is run on Alpha.
  • WHAT: Research compute jobs can be of any CPU/GPU-scale and duration, given the hardware specifications, guidelines, and context below.
  • WHY EARLY ADOPTERS: Alpha will be in start-up mode during its initial months of operation, meaning early users will be expected to help work out configuration issues, and ensure the necessary software is installed.
Here are some further guidelines and context:
  • Empire AI will allocate resources for this initial phase of Alpha following these operating principles: To achieve equity of usage across the six academic systems sharing the instrument, and to maximize utilization of Alpha’s capacity.
  • Priority will be given to projects that best harness the capabilities of Alpha.
  • Depending on the total number of work orders submitted by any one academic institution and the distribution of work orders submitted among institutions, Empire AI may ask a university to prioritize its corresponding work order requests, following whatever procedure that academic institution chooses.
  • When considering appropriate research data for use of Alpha, note that Alpha is neither HIPAA nor NIST-800-171 compliant.
  • Empire AI expects to make Alpha available in this fashion for about one year (i.e., through approximately November 2025), subject to change with notice to the user community.
Consider this call for Empire AI WORs as just the first one. There may be subsequent calls as we gain more experience in using Alpha after this start-up phase. The process for soliciting work orders may change as well.

Finally, it is through the generosity of the Simons Foundation and its Flatiron Institute that Empire AI can so rapidly provide access to an initial instrument for our shared research use.  Through these means, there are no institutional or user fees for running compute jobs on Alpha.  We are piloting Alpha to understand user demand, usage patterns, classes of users, and types of jobs and workloads. Empire AI and Stony Brook University will use our collective experience with Alpha to plan user fees and resource allocation for future builds of Empire AI.

If you have any questions, please contact Robert Harrison at Robert.Harrison@stonybrook.edu.

Please consider submitting a work order to take advantage of this opportunity. I sincerely hope that Stony Brook alone overwhelms the Alpha system, but in any case it will help us understand the need for this and additional GPU resources.

Kevin Gardner
Vice President for Research
Stony Brook University
kevin.gardner@stonybrook.edu
O: 631.632.7932 | M: 603.767.4654 |


Empire AI: Alpha Hardware Specifications

12 HGX Nodes

  • 8 H100 80GB GPUs
  • 10 400Gb/s ConnectX-7 NIC Cards (8 for IB and 2 for Ethernet)
  • 30TB NVMe caching space
  • 2TB of system memory
Non-blocking NDR fabric cabled for rail configuration
  • 8 network switches, 96 optical connections
4 service nodes
  • 2 login nodes and 2 cluster management nodes (NVIDIA Base Command with licenses for all gear)
2PB of DDN Storage
  • 4 x 720TB Flash storage (home directories, training data, snap shots)
*Empire AI Consortium, Inc.’s purpose as a non-profit corporation is to develop and oversee a shared high performance computing facility that will promote responsible research and development, including the advancement of the ethical and public interest uses of artificial intelligence technologies in New York.

How to Tell the Chat GPT: Privacy Please!

a privacy sign hanging from a door knob

 

To opt out of using your data to train Chat GPT:

Go to Settings

image of settings location

Click on Data Controls

image of settings menu

Click on “Improve the model for everyone” and set it to off.

Image of the prompt that asks you if you really want to not help the model train.

ALTERNATIVELY:

You can go to this website, and click on the button in the upper right that says “Make a Privacy Request”

Note that Enterprise ChatGPT has privacy enabled by default.

“You can access the privacy portal for ChatGPT services by visiting OpenAI’s Privacy Policy page. This page provides information on how your data is handled, including collection, usage, and your rights regarding your information. If you’re looking for specific privacy settings or options to manage your data, they are generally outlined in the policy itself or via account settings within the OpenAI platform.”

 

image of the options inside the privacy portal.

Here you can also Request a copy of the data, delete your account and make a ChatGPT Personal Data Removal requests.

You should absolutely avail yourself of these settings if you are using this for work where you are loading up sensitive information. We work with FERPA (Family Educational Rights and Privacy Act) data, HIPAA (Health Insurance Portability and Accountability Act) data and research data here at the university and need to be careful.

 

Services for businesses, such as ChatGPT Team, ChatGPT Enterprise, and our API Platform

“By default, we do not train on any inputs or outputs from our products for business users, including ChatGPT Team, ChatGPT Enterprise, and the API. We offer API customers a way to opt-in to share data with us, such as by providing feedback in the Playground, which we then use to improve our models. Unless they explicitly opt-in, organizations are opted out of data-sharing by default.

Please see our Enterprise Privacy page for information on how we handle business data.”

 

 

 

AI still wowed me today

So – if you follow what I write, you know that I have a little obsession with what AI can do with just your course syllabus.   Well check this out:

I went to https://notebooklm.google.com/

Then I pasted the text from my colleague’s syllabus, entered the Notebook Guide and asked it for an Audio Overview.   That is all.

 

Mind Blown.

 

Adobe Express does a little bit of everything

Without even bothering to explore Adobe Express much, I just assumed it was a dumbed down version of Adobe Photoshop.  It does satisfy that description, but it is also a dumbed down Premier, InDesign, Illustrator, Acrobat, converter, AND text to graphic AI generator.

 

a screenshot of the menu items under quick actions in Adobe Express: remove background, resize image, logo maker, convert to gif, convert to pdf, generate QR Code, convert to svg, convert from pdf, text to image

menu of quick actions

another screenshot of the menu items under quick actions in Adobe Express: text effects, collage maker, convert to png, animate from audio, trim video, caption video, merge videos, crop video, convert to mp4, crop image, edit pdf, combine files, organize pages

part two of quick action items in adobe express

QR Code maker:

Generative AI image style:

What is a seawolf?

Maybe make it look more like a university mascot?

 

or start with a real photo and just mess with that using AI

That one “Animate from AudioI’ve blogged about before: https://you.stonybrook.edu/jadams/2024/06/12/adobe-education-institute-animate-from-audio-in-adobe-express/

 

Here is that Collage Tool:

Teaching and Learning with AI Conference: My Poster Presentation

It isn’t really a poster presentation – click for the full slide deck

click here for presentation.

click here for presentation.

The general concept of this presentation is to remind instructors that by creating a solid syllabus, they have already done most of the work that will allow a generative AI tool to be extremely helpful to both the students and to themselves.

DuckDuckGo’s new private way to use AI

Anonymous access to popular AI models, including GPT-3.5, Claude 3, and open-source Llama 3 and Mixtral.

 

click on image to go to chat interface

 

 

 

‘DuckDuckGo AI Chat is an anonymous way to access popular AI chatbots – currently, Open AI’s GPT 3.5 Turbo, Anthropic’s Claude 3 Haiku, and two open-source models (Meta Llama 3 and Mistral’s Mixtral 8x7B), with more to come. This optional feature is free to use within a daily limit, and can easily be switched off.

  • Chats are private, anonymized by us, and are not used for any AI model training.
  • Find DuckDuckGo AI Chat at duck.aiduckduckgo.com/chat, on your search results page under the Chat tab, or via the !ai and !chat bang shortcuts. They all take you to the same place.
  • Improvements are already on the way. Our roadmap includes adding more chat models and browser entry points. We’re also exploring a paid plan for access to higher daily usage limits and more advanced models.”

 

Read more here

Today’s Google Developer Keynote livestream – starts at 1pm

OMG – too many names!

Generative Voice AI

Yet another toy to place with.  Educate yourself about it as well.  The Speech to Speech option can be used to do a deepfake, but for the voice.

https://elevenlabs.io/

image of voice generator from website

 

FACT2 AI Faculty Development Workshops 2024

Please check out the Spring 2024 Workshops below. The workshops will be offered at no cost, and will be recorded, and the recordings will be available on the Playlist below.

Getting Started with Artificial Intelligence

 

  • March 12, 2024, 10:00 AM – 11:00 AM
  • March 15, 2024, 1:00 PM – 2:00 PM

 

Description: As GPT has gained widespread attention since the release of GPT 3.5 in late 2022, our digital landscape has been evolving quickly.  With academia, media, governments, and corporations focused on the furthest reaches of artificial intelligence, it is easy to get overwhelmed by the possibilities this powerful tool offers us. This hands-on session focuses not on the outer limits of AI, but instead on:

 

  • Getting started with GPT/ Understanding its roots
  • Establishing best practices for general usage
  • Expanding our comfort zone
  • Working with AI to determine how it can and cannot meet our individual needs Attendees will have the option to observe, click along with provided usage examples, and share their results with others.

 

Presenter: Robert Becker


Exploring Potential and Pitfalls of AI Use in the Classroom 

 

  • March 19, 2024, 10:00 AM – 11:00 AM
  • March 22, 2024, 1:00 PM – 2:00 PM

 

Description: AI tools have access to enormous data and use enormous processing power to generate plausible patterns that can save time, offer sophisticated text for users to consider, and boost thinking/learning and writing/communication for students and faculty alike. But they also have glaring weaknesses, such as their inability to recognize reality (versus merely “plausible” strings of words), understand context or culture, offer unbiased and ethical responses, avoid privacy or security infringement, etc. Join this hands-on session to learn and share how to help students recognize pitfalls of AI, as well as explore its potentials. We will collaboratively develop and exchange learning activities for our students.

 

Presenter: Shyam Sharma, Michael Murphy, & Cynthia Davidson


Developing Syllabus Statements on AI Use

 

  • March 26, 2024, 10:00 AM – 11:00 AM
  • March 29, 2024, 1:00 PM – 2:00 PM

 

Description: As faculty navigate the use of Chat GPT and other AI tools in their courses, it is important to communicate their expectations with the students.  This session will provide participants with examples of syllabus statements being used within SUNY and beyond.  The examples will reflect varying levels of permitted AI usage:

 

  • Minimal/ none
  • Some use with specific rules/ circumstances
  • Expected/ required

Participants will be encouraged to discuss the examples and how they relate to their own current or future AI use in their courses.  Following an open discussion and sharing of ideas, individuals will collaborate to adapt/ develop a syllabus statement based on their desired level of AI usage.  Participants will leave the workshop with a syllabus statement that they may use in their course.
 

Presenter: Meghanne Freivald & Keith Landa


AI As An Assistant

 

  • April 2, 2024, 10:00 AM – 11:00 AM
  • April 5, 2024, 1:00 PM – 2:00 PM

 

Description: AI can be helpful for a number of tasks for students, including suggesting writing improvements, summarizing or rephrasing concepts they are learning, helping with ideation to overcome “blank page syndrome”, or even generating questions for review. The commonly available tools, such as ChatGPT, have been trained on a vast corpus of information that covers many subjects, albeit non-uniformly, and its ability to provide accurate and helpful information varies by subject, as it its likelihood that it will provide incorrect information (they call this a “hallucination”).

Those are factors that are out of your control, but you can mitigate them….with a well crafted prompt.

This workshop aims to help you assess how effective these tools could be for your students – looking at whether they work out of the box, and whether the right prompt can address gaps and concerns. Even if you find that these tools are really not well suited for your course, that can be something that you can share with your students and have a chance to guide them to other options.

Presenter: Maureen Larsen & Brian Cepuran


AI Tools to Help You Build Your Course

 

  • April 9, 2024, 10:00 AM – 11:00 AM
  • April 12, 2024, 1:00 PM – 2:00 PM

 

Description: There are a variety of tools to help instructors build out their courses, from designing learning outcomes, to creating course content, to creating quizzes and other assessments.  Both general-purpose and specialized AI tools will be demonstrated in the workshop, and participants will have opportunities to try them out.  Both faculty and the instructional support staff that assist them are encouraged to attend. No prior AI experience required.

Presenter: Brian Cepuran & Keith Landa

Register Now