AI Workshop: Missing, Improved, Removed

TODO: collapasable, copyable, download, etc

The following is an outline for a "Using AI Workshop" designed for high school teachers. Please analyze it and give a report on what is missing, what can be improved, and what should be removed. For missing and improved sections, please output the next outline.

• Who AM I, my journey
  • Background, age, roughly similar background as the staff, tell my story, connect and ground
• Introduction (the issue, pros/cons)
  • AI Metafor - Fire, Splitting the Atom, AI
    • Revolutionary Impact
    • Tool vs Threat
    • Requires Learning and Mastery
    • Propagation
  • AI Metafor - clay with indentations - mold of the world
  • AI is a college intern that you can give low level tasks and they may or may not do a good job on them
    • Super smart, but also lacks common sense
  • Limitations of today's AI Tools
    • ChatGPT -
      • Hallucinations
      • Safety measures that limit usage
    • Midjourney
      • Hands, eyes, fine details
      • It has a hard time creating images it has never seen itself
• Example of Good, Bad, and Ugly outcomes of AI
  • Good
    • Average people can use their personal "AI team" to accomplish exceptional things
      • Its possible that AI will help us solve energy, climate change, poverty
  • Bad
    • LLM are trained on human's creative work, so there are moral and legal copyright risks that we are stealing from others (greg rutkowski artwork)
    • A lot of what we fear from AI is already here and it is killing us. It's called capitolism.
      • wealth inequality, labor exploitation, optimizing for the wrong priorities
      • AI is a force multiplier to capitolism.
    • AI could destroy entry level jobs and expert level jobs
      • In the short term, physical labor could be safe
      • Until AI makes major breakthroughs in material science, robotics
    • Universal Basic Income will save us!
      • Except US would never allow it, look at homeless crisis
  • Ugly
    • AI can be demotivating and de-humanizing
    • Right now ChatGPT is more knowledgable than any human on earth. It makes some basic logic mistakes, it hallucentates and gives wrong answers, it...
      • AI could erode the very nature of truth and unravel the fabric of society
        • Just acceleating a current authoratarian trend
    • Future versions will be expert level at most tasks. It could formulate a plan and put it into action that kills a majority of humans, before we're even aware
    • Saftey and Alignmnent - This focus is on how to contain AI and how to align it's goals and priorities with humankind.
• Breakout Room #1 (Writing and Critical Thought)
  • When a student turns in work that is clearly created by AI, how do you react?
  • Old World Assignment: Read this article and write up your reaction.
    • New World Assignment: Asseble an AI team and together read the article. Together Analyze it. Together, write up your augmented reaction.
  • What ways could you challenge a student to go further?
    • Analyze the homework inputs and the student's outputs using AI
      • As AI what questions the homework assignment didn't ask?
      • Give an AI the studnent's answer and ask them what are the weakest points of their argument
      • Can you give me an outline of topics that I should cover, which I did nto?
      • Can you give me a prioritized list of where I should go deeper in my answer?
      • What AI personas could a student invent to build their "dream team" to produce the best outcome for a homework assignment?
  • What are ways that students could "show their work"?
    • What if you had students mark up prompts vs output vs edits?
    • What if you had students record iterations, versions, and progress of an answer?
    • How do you evaluate and grade group projects?
      • What if the group project was one student and 3 AIs?
• Breakout Room #2 (Visual Arts and Creativity)
  • Artists have always sought examples, inspiration, and raw building materials for making something new and better
  • Generative AI is really terrible at creating images of something that it hasn't ever seen before. But it is really good at generating variations that might provoke creativity in an artist.
  • Generative AI cannot take too much detail or instruction in one prompt or it gets confused
  • Take a problems or goal and decompose the pieces - foreground, background, character, setting, mood, etc and generate images for each peice. Then put all of those pieces together into a mood board for creating your own work from scratch.
• Preparing our students for an indeterminant future
  • AI Literacy
    • It is more than a simple chat message and accepting the first answer
    • Prompt Engineering
    • Picking the best AI tool(s)
      • Different text models (GPT, Claude, Bard)
      • Different image models (Midjourney, StableDiffusion, Adobe Firefly)
    • Understanding the foundation
      • Types of AI (Large Language Models with a Transformer Architecture)
    • Different techniques to train and fine-tune models
      • LORA
    • Computer literacy and programing
      • Never a bad thing to learn some scripting
      • Automatically taking the outputs from AI teammember 1 and giving them as inputs to AI team members 2 and 3
      • Creative possibilities are unlimited
  • Characteristics to cultivate in our students
    • Curiosity (What about X? How does Y even work?)
    • Skepticism (The chatbot said X, but that doesn't seem likely, I'm calling B.S and asking for proof)
    • Design and System's thinking (I should create a template and a checklist for using my AI Team)
• Q&A This is an excellent outline for a "Using AI Workshop" for high school teachers. The topics are relevant and comprehensive, covering both theoretical and practical aspects of AI, its implications for society, and how it can be used in education.

However, some sections could be improved or added to make the workshop even more effective. Here's the new outline with recommended improvements:

• Who AM I, my journey
  • Background, age, roughly similar background as the staff, tell my story, connect and ground
• Introduction (the issue, pros/cons)
  • The History of AI
    • Brief explanation of AI's evolution and its current state
  • AI Metafor - Fire, Splitting the Atom, AI
    • Revolutionary Impact
    • Tool vs Threat
    • Requires Learning and Mastery
    • Propagation
  • AI Metafor - clay with indentations - mold of the world
  • AI as a tool and its practical applications
    • How AI is currently used in different sectors, such as healthcare, finance, and education.
  • AI as a college intern that you can give low level tasks and they may or may not do a good job on them
    • Super smart, but also lacks common sense
  • Limitations and ethics of today's AI Tools
    • Include more about ethical considerations, data privacy, and bias issues
    • Case studies or examples of ethical dilemmas in AI usage
  • Breakout Room #0 (Ethical Considerations)
    • Discussion about how to teach ethics of AI usage to students
    • Brainstorming on how to integrate ethical considerations into AI lessons
• Example of Good, Bad, and Ugly outcomes of AI
  • This section seems comprehensive. Keep it as is.
• Breakout Room #1 (Writing and Critical Thought)
  • This section is excellent. However, it might be beneficial to provide more examples of AI-assisted assignments for different subjects, not just English or Humanities.
  • Add exercises to practice the application of AI in the classroom.
• Breakout Room #2 (Visual Arts and Creativity)
  • This section is well done but could include more examples of how generative AI can be used in creative projects or lessons.
  • Add exercises to practice the application of AI in arts and creativity.
• Preparing our students for an indeterminant future
  • This section is well thought out. However, add some real-world examples or case studies of successful AI applications and include activities to design possible future scenarios involving AI.
• Q&A
  • It might be useful to split the Q&A section based on the different themes of the workshop to ensure that all areas are covered.
  • Consider preparing some common questions and answers ahead of time to ensure you're ready for any queries that arise.

The only element that could potentially be removed or revised is the controversial statement about capitalism. This statement may cause unnecessary political debates in a setting that should primarily focus on AI. However, if this issue is relevant to your audience, it might be worth keeping, but framed more neutrally to foster a productive discussion rather than a debate.I can't tell which parts you changed. Can you output that again and bold items you added and strike through items you deleted or changed?Sure, here is your revised outline with bolded additions and struck through changes.

• Who AM I, my journey
  • Background, age, roughly similar background as the staff, tell my story, connect and ground
• Introduction (the issue, pros/cons)
  • The History of AI
    • Brief explanation of AI's evolution and its current state
  • AI Metafor - Fire, Splitting the Atom, AI
    • Revolutionary Impact
    • Tool vs Threat
    • Requires Learning and Mastery
    • Propagation
  • AI Metafor - clay with indentations - mold of the world
  • AI as a tool and its practical applications
    • How AI is currently used in different sectors, such as healthcare, finance, and education
  • AI as a college intern that you can give low level tasks and they may or may not do a good job on them
    • Super smart, but also lacks common sense
  • Limitations and ethics of today's AI Tools
    • Include more about ethical considerations, data privacy, and bias issues
    • Case studies or examples of ethical dilemmas in AI usage
  • Breakout Room #0 (Ethical Considerations)
    • Discussion about how to teach ethics of AI usage to students
    • Brainstorming on how to integrate ethical considerations into AI lessons
• Example of Good, Bad, and Ugly outcomes of AI
  • A lot of what we fear from AI is already here and it is killing us. It's called capitalism.
• Breakout Room #1 (Writing and Critical Thought)
  • Add exercises to practice the application of AI in the classroom
• Breakout Room #2 (Visual Arts and Creativity)
  • Add exercises to practice the application of AI in arts and creativity
• Preparing our students for an indeterminant future
  • Add some real-world examples or case studies of successful AI applications and include activities to design possible future scenarios involving AI
• Q&A
  • Split the Q&A section based on the different themes of the workshop to ensure that all areas are covered
  • Consider preparing some common questions and answers ahead of time to ensure you're ready for any queries that arise What was the year and first major breakthrough of large language models, please answer brieflyThe first major breakthrough of large language models was in 2018 with the introduction of GPT (Generative Pretrained Transformer) by OpenAI. The transformer architecture that GPT used, which was first introduced in the "Attention is All You Need" paper by Vaswani et al. in 2017, allowed for a significant increase in the size and capabilities of language models. GPT was a departure from previous models as it was trained to predict the next word in a sentence, making it more capable of generating human-like text.Was there a major breakthrough a decade or so earlier around neural networks?Yes, there was a significant breakthrough in the field of neural networks about a decade before the advent of large language models. This was the development of deep learning techniques, specifically the creation and successful training of deep neural networks.

A key event was in 2006-2007 when Geoffrey Hinton and his team demonstrated that deep belief networks, a kind of deep neural network, could be effectively trained using a method called greedy layer-wise pre-training, overcoming earlier problems with training deep networks.

Another significant breakthrough occurred in 2012 with the application of convolutional neural networks (CNNs) to image classification in the ImageNet competition. A team led by Alex Krizhevsky, Ilya Sutskever, and Geoffrey Hinton developed a model (often referred to as "AlexNet") that significantly outperformed previous state-of-the-art models on the ImageNet image classification task. This marked a turning point in the field, sparking renewed interest in neural networks and their potential capabilities.What year was alphago?AlphaGo, developed by Google's DeepMind, made headlines in 2016 when it defeated the world champion Go player, Lee Sedol, in a five-game match. This was a significant milestone because the complexity of the game Go is extremely high, with more potential board configurations than there are atoms in the universe, making it a challenge that classical AI approaches could not effectively tackle. AlphaGo's success demonstrated the potential of AI systems based on deep learning and reinforced learning.Does AlphaGo have anything to do with LLMs?AlphaGo, developed by DeepMind, and Large Language Models (LLMs) like GPT-3, developed by OpenAI, are both groundbreaking AI technologies, but they have different functions and architectures.

AlphaGo is specialized for the game of Go and uses a combination of Monte Carlo tree search (MCTS) and deep neural networks trained with reinforcement learning and supervised learning.

On the other hand, LLMs are general-purpose models that use a transformer neural network architecture and are trained on a large corpus of text data. LLMs are designed to understand and generate human-like text, making them useful for a wide range of natural language processing tasks.

While both AlphaGo and LLMs are powerful tools in the field of artificial intelligence, they serve different purposes and use different techniques. As of my knowledge cutoff in September 2021, there wasn't a direct link or application between AlphaGo's architecture and LLMs, though advancements and techniques in one area of AI often influence and inspire developments in other areas. In 2017 with attention is all you need, what was the famous examples or tool?The "Attention is All You Need" paper, published in 2017, introduced the concept of the "Transformer" model, which has since revolutionized the field of natural language processing (NLP).

The main innovation of the Transformer model was its use of self-attention mechanisms, which allowed the model to weigh the importance of different words in a sentence when generating a prediction. This provided a way to handle long-range dependencies in language and led to a dramatic improvement in performance on a range of NLP tasks.

One of the most famous applications of the Transformer model is OpenAI's GPT (Generative Pretrained Transformer) series. GPT-1 was first introduced in 2018, and by 2021 (the time of my last training data), the latest version was GPT-3. These models, trained on a vast corpus of internet text, have shown remarkable proficiency in generating human-like text and have been used in a wide variety of applications, from writing essays to coding assistance to creating poetry.Can you give a brief history with years of Dall-e, Midjourney, and Stable Diffusion?As of my last update in September 2021, here's the information available:

• DALL-E: This model was introduced by OpenAI in January 2021. It's essentially a variant of the GPT-3 model, trained to generate images from textual descriptions. It has the ability to produce highly original -- even surreal -- images from prompts that combine unrelated concepts.

Regarding Midjourney and Stable Diffusion, as of my last training cut-off in September 2021, there's no specific information or models with those names that have been publicly released or widely recognized in the AI community. This could be due to the fact that these models were released or gained significance after my last update. You may need to refer to the most recent sources or directly contact the relevant organizations for the most accurate and up-to-date information.

AI Workshop: Missing, Improved, Removed created at Tue Jan 20 1970 13:23:03 GMT+0000 (Coordinated Universal Time)