Review from last week (the Cloud)
Reminder Discussion Board
Presentation
Discussion and Hands on
Review the lesson following the class
Lots of fun things to try: Ask a question of ChatGPT, Come up with a vacation agenda with Claude, Use notebook to organize your classes, Ask Pi how to handle an overanxious puppy. Let us know how it goes!
The Zoom classroom:
https://sdccd-edu.zoom.us/j/9191959460?pwd=OXh0RE9ZTVZTWElTMUQ0ZzAxQzExdz09
Passcode: emeritus
· Introduce AI (Artificial Intelligence) and highlight how it is already being used
· Introduce the terms of AI (machine learning, deep learning reinforcement learning)
· Discuss techniques used (Natural Language Processing, Large Language Models and introduces prompts
· Describe two specific applications of AI: Google Lens and ChatGPT
· Introduce three additional tools as time permits: Pi, NotebookLM (Google) and Claude
· Information on privacy and ethical use
Artificial Intelligence (AI) refers to computer systems that perform tasks typically requiring human intelligence. These tasks include learning, reasoning, problem-solving, understanding natural language, speech recognition, and visual perception. AI technologies use algorithms and massive amounts of data to learn from patterns and features in the data. AI can be categorized into narrow AI, which is designed to perform a narrow task (e.g., facial recognition or internet searches), and general AI, which would outperform humans at nearly every cognitive task. It's applied in various fields, including healthcare, finance, autonomous vehicles, and customer service, enhancing efficiency, precision, and automating processes. You are already using AI in different ways:
· Smartphones and virtual assistants: When you ask Siri or Alexa a question, they understand and respond. AI can even distinguish between two voices (Our Alexa device recognizes us as Kim and Mary).
· Email filters: AI is used to filter out spam emails. Over time, it learns what is junk mail and what is important. It learns this with your own interactions as well as using algorithms to predict spam.
· Navigation apps: AI is used to provide the best driving directions, analyzing traffic patterns, road closures and even the time of day to suggest the best route.
· Online shopping and recommendations: When shopping online, websites may offer recommended products based on what you have ordered before or what others order when they purchase this item.
· Healthcare devices: Products like heart monitors and wearable fitness trackers use AI to analyze things like heart rate and activity. In some cases, they may be able to alert you to potential issues.
· Social media: Sites like Facebook use AI to show you things that you might be interested in, based on what you have interacted with before. They use your connections and your likes to predict who you might want to follow. And of course, they offer friend suggestions based on friends you have (pulling in data you have provided such as education history, previous locations, pets, interests and more).
Using AI is like teaching a computer to think and learn. Imagine AI as a smart robot that can do tasks, solve problems, and even talk like us, depending on what it has been taught.
AI uses vast amounts of data, recognizing patterns and making decisions. It includes the use of:
· Machine learning (where the system learns from the data it is given)
· Deep learning (where the system learns in an unsupervised way from unstructured data, learning to use predictions to provide answers)
· Reinforcement learning (where the system makes decisions based on actions and is evaluated along the way).
Supervised (with labels, used for simple chatbots) and Unsupervised (all data thrown together)
Data is extracted from this set and fed into an algorithm which will then group the objects into similar patterns and predict a solution.
Machine learning is used to create deep learning (neural networks) and reinforcement learning (evaluating predictions)
Machine learning: Show pictures of different fruits. Class “trains” the machine by labeling a set of fruits. (apples, bananas, oranges). Then, show new pictures and ask them to “predict” which fruit it is by recognizing patterns (shape, color). Takeaway message: Once the machine knows that an apple and banana look like, it can sort new fruits based on past learning.
Deep learning: Deep learning relies on layers working together. The layers are past of a large set of “unlabeled” data. These layers work together to decide. This activity will learn to recognize patterns at different levels using the annote tool in Zoom.
· Step 1: Draw a simple shape (square, triangle)
· Step 2: Think of each shape as a “feature” the model is trying to learn.
· Step 3: Add more details to the drawing and explain that as the model becomes more complex (more layers), it can recognize more intricate patterns
Takeaway message: Deep learning starts with simple features (like the edges of a shape) and gradually learns more complex details (like feature and extra details).
Reinforcement learning: You probably have used reinforcement learning in the past, whether training your pet to sit or potty training your child. In both examples, we provide a reward if the behavior is what we were looking for. After a while, the pet or child realize that if they get a reward when they do this behavior, they will most likely repeat the action. Takeaway message: Reinforcement learning involves learning through rewards and punishments, similar to the pet and child example.
Ai uses a series of steps to train the machine for AI. They include:
Data Collection: Gather large datasets of text from various sources to cover a broad spectrum of language use.
Preprocessing: Clean and organize the data
Model Selection: Choose an appropriate machine learning model (ChatGPT? DALL-E?)
Training: Use the prepared dataset to train the model, adjusting it to minimize errors like language understanding, translation, or text generation.
Evaluation: Test the model on unseen data to evaluate its performance and ability to understand and generate language accurately.
Fine-tuning: Adjust the model based on performance feedback to improve its understanding and communication abilities.
Deployment: Implement the trained model in real-world applications for tasks like chatbots, voice assistants, or text analysis tools.
Language is the key to teaching AI. We teach AI by showing it lots of text, which helps it learn to make sense of words and sentences like a human does. This way, AI can understand and respond to us better. As we keep teaching AI with more language, it gets smarter and can talk or write in ways that are more helpful and interesting.
There are two important concepts involving AI training which include language. One is the method used to allow computers to understand and interact with our language (Natural Language Processing) and the other is a model developed to perform more complex language tasks (large language model).
Natural Language processing (NLP) is the interaction between computers and humans through language. It involves tasks like text understanding, language generation, analyzing sentiments, and more. Some of the NLP techniques include rule-based approaches (chatbots from SDG&E), statistical methods (analyzing data) and machine learning algorithms. Your virtual assistants like Alexa use NLP.
Some points about NLP:
• Supervised learning, with labeled text (only works from data it is given)
• Allows machines to interpret, translate, summarize, and analyze
• Can produce summaries (“Alexa, play my favorite music”)
• Can create classifications (From SDGE, the chatbot follows a flowchart to arrive at an answer just for you)
Large Language Models (LLMs) are powerful tools created using NLP. They can read, understand, and even write text that sounds like it was written by a human. They are neural network-based models (think of your brain with all of the connections: neural networks have a similar design) which can be used to understand and generate human-like text. ChatGPT is an example of a LLM and is a more specific tool for generating assistance than Alexa.
Some points about LLMs:
• Uses unsupervised learning (it is fed lots and lots of data from a wide variety of sources) as well as supervised
• It can then use that data to understand and generate human-like content (images, videos)
• Because it uses a variety of information, it is more adept at understanding semantics, context and syntax
Prompts are the instructions or questions asked of an AI like ChatGPT to get a response (or output). The prompt can be a question, a statement, or a command. Simple prompts will create simple answers, but a thoughtful prompt may produce a much more meaningful example. Here are some things to keep in mind with prompts:
· Clarity: Be as clear and specific as possible
· Context: Providing context helps the AI to understand the background, especially important for complex topics
· Purpose: Define what you want from the AI. Is it information, an opinion, a creative output, a calculation or whatever.
· Type of question to ask can include direct questions (facts or specific information),open-ended requests (for creative or opinion-based responses) or commands (translating, solving a problem or generating an image)
· Other tips:
o Tell the AI about your audience so the answer can be directed to them
o Sequence the prompts: start basic and gradually build up to it with follow-up prompts
o Use examples to guide AI (design an image of a cat in the style of Van Gogh)
· Local History Inquiry: "Can you tell me about the history of La Mesa?”
· Recipe Requests: "How do I make a traditional apple pie?"
· Problem solving: “How do I fix a leaky faucet?”
· Gardening Tips: "What are some tips for growing tomatoes in a home garden?"
· Health Tips for Seniors: "What are some healthy lifestyle tips for seniors?"
· Book Recommendations: "Can you recommend some books for someone who loves mystery novels?"
· Travel Information: "What are some senior-friendly travel destinations in Europe?" Or even: “I would like to plan a trip to Minnesota…”
· Technology Help: "Can you explain how to use email on a smartphone?"
· Memory Lane: "What were the popular songs or movies in the 1960s?"
· Language Learning: "I want to learn basic Spanish. Can you help?"
· Brain Teasers and Puzzles: "Can you give me a fun crossword puzzle or a brain teaser?"
· Storytelling: "Can you create a short story about a grandparent's adventure?"
Suggestion: Start with a simple prompt "What is the iCloud". See how your AI answers
Next, try a detailed prompt. I used the following and came up with some excellent ideas:
As an instructor in Technology for older adults, I am trying to assist my older adults who may be struggling with understanding the iCloud. You will help by responding as an Apple tech support expert to help older adults better understand iCloud. Can you explain the concept of the iCloud in respect to storing photos and how it would affect their storage on their device? Include examples which will be easy for an older adult to grasp.
Privacy: Chats are saved by default for training the chatbot. And since these tools use large language models, it would also seem that they are using your data to build skills. So assume that anything that you request will become part of the data used to train this app.
You can turn off the saving of chats in settings>data controls>turn off chat history and training.
You will have to do this across browsers. You can also delete your account using the same location in settings.
If you disable history, new chats will disappear within 30 days Read about privacy on OpenAI:
Ethical use: There are ethical questions in using AI. It often lacks accountability and may be subject to bias. It may also contribute to social engineering, misinformation and even addiction.
Always verify before posting
Keep in mind that the models are biased and might not reflect everyone
In cases where detecting the use of AI is important (such as education), there are detection programs like gptzero. https://gptzero.me/
An AI application is a tool or system that uses artificial intelligence to perform tasks usually requiring human intelligence. This can include recognizing speech, making decisions, translating languages, or identifying objects in images. Some examples of AI applications that you are currently using include things like virtual assistants, facial recognition, email filtering, customer service chatbots, health and fitness trackers and more.
Google lens: You may already be familiar with Google Lens. Google lens allows us to upload a picture or take a picture and AI recognizes objects, landmarks, animals and more. Its image recognition uses AI to identify and provide information about objects in the photos. It then uses AI to analyze and interpret images from its database.
Try it: Open Google Lens on your device (either through the Google search icon on your smartphone or by going to lens.google.com). Take or upload a simple image. Note that Google Lens identifies the object and offers other related information (type of object, related objects, shopping links). AI compares the image to similar objects it has learned from its database. This type of image recognition is also used in security, shopping apps and navigation.
ChatGPT: ChatGPT is an OpenAI product which generates text based on a prompt. One way to use this is to predict what comes next in a sentence based on learned patterns from text. This is the type of technology that is included in “suggestions” in emails or chatbots.
The basic version of ChatGPT is free and accessible to everyone, offering features like text-based conversation and general query resolution. For users seeking enhanced capabilities, such as faster responses, website analysis, image creation, and more advanced options, there is a premium version available at $20/month.
Try it: Go to https://chatgpt.com . No need to sign in just yet. Enter a simple prompt: “Once upon a time…” or “Once I came upon…” or any beginning to a sentence. Type it into the box and click on enter. You will see ChatGPT generate the text. This demonstrates how text generation works and helps users to generate meaningful content.
1. Ask it to complete or continue a story based on your input. For example, “Once upon a time…” See how it looks!
2. Ask it for a recipe: Type “I have chicken, potatoes and spinach. Can you suggest a recipe?”
3. Ask for weekend information: “I have grandkids visiting this weekend. What are some free, kid-friendly events in San Diego that they might enjoy?”
4. Ask for help with an email: “Help me write an email to my friend asking them to dinner this weekend”
5. Ask a tech related question: “ How do I delete files from the iCloud to create more storage space?”
Pi: Pi is a companion virtual assistant. You can download the app from the Apple or the Google Play store. You can then set it up to respond to you verbally while you ask it questions. It carries on the conversation and remembers things. You can access PI by going to: https://pi.ai/discover We will demonstrate Pi and the conversational abilities that it has.
NotebookLM (Google): This brand-new tool takes information that you have and creates guides for you, such as a study guide, table of contents timeline. It can also use your data to answer questions that you have and can even create a “podcast” on what was provided. You will need a Google email to access it. Go to: notebooklm.google.com. Select New Guide. Upload any files you have. Let’s view a notebook guide that I created in all of my sources with AI. How can you use this tool? Here are some suggestions:
· Travel planning: Create a notebook for travel including destination research, itinerary planning, travel tips, packing lists, etc.
· Lifelong learning: Upload your handouts from a class and use this tool to generate study guides, summaries and to ask qualifying questions
· Recipe collection and meal planning: Upload your favorite recipes and cooking tips. You can then organize recipes, suggest meal ideas and even generate shopping lists
· Book clubs: Upload book summaries, reviews and personal reading notes. Use NotebookLM to create discussion questions and personalized reading lists
Claude: You will have to sign in with an email account. Claude is an AI assistant which can help with a variety of tasks, including answering questions, analyzing things and even problem solving. If you are confused on a concept, it can explain it in several ways to make it more understandable. It does not retain personal information from conversations.
An example would be for Claude to plan a healthy meal which would include varieties of anti-inflammatory foods.
· Request: “I need help in designing a meal which would target my inflammation. I would like this meal to include foods which can decrease inflammation. These foods should be easy to find in the store. I like simple meals, and would prefer either vegan or fish.
· Results: You will see a number of options and what ingredients are anti-inflammatory.
· Next: Use one of the meals and ask for a recipe.
· Then: Request a shopping list for the meal along with any sides that you want.
Consider backing up with a request for a shopping list!
· Help with technology-related questions ( how to use a new smartphone or set up video calls with family)
· Assist with researching health topics, explaining medical terms, or providing tips for managing chronic conditions
· Planning travel itineraries, suggest age-friendly activities, or recommend books and movies
· Aid in writing emails or letters, proofreading documents, or even learning new skills like basic coding or digital art
· Provide discussions on history, current events, or provide brain teasers and puzzles
· Tips on gardening, home organization, or simple recipes tailored to specific dietary needs.
Resources for Further Learning:
Learn about ChatGPT
Or, use the tool itself to find more. In this example, I went to Pi to carry on a conversation about Pi. Here is how it looked. Each segment begins with a question from me and the response from Pi. Responses are in italics
How to explain Pi to my older adult class:
Pi is an AI (Artificial Intelligence) that can carry on human-like conversations. You can chat with Pi just like you would with a person - ask questions, share your thoughts, and even crack a few jokes! Pi is designed to understand and respond to a wide variety of topics, from cooking recipes to current events. It's like having a knowledgeable friend at your fingertips!
How does it work?
Pi works by using advanced natural language processing algorithms and machine learning techniques. Here's a brief breakdown:
When you send a message to Pi, it processes and analyzes the text to understand the meaning behind it.
Pi then searches its vast knowledge base for relevant information and generates a response based on that information.
Over time, Pi learns from your conversations and adapts its responses to better suit your preferences and interests.
The result is a personalized chat experience that feels natural and engaging.
In short, Pi is a complex software program that can communicate with people in a way that feels human-like. It's pretty amazing!
Does it save my conversations?
Yes, Pi does save your conversations for a few reasons:
To provide a more personalized experience. By remembering your previous conversations, Pi can tailor its responses to better fit your interests and needs.
To improve its performance over time. By analyzing past conversations, Pi can learn and adapt to better communicate with users like you.
To ensure safety and compliance. In rare cases, conversations may be reviewed to ensure that Pi is adhering to ethical and legal guidelines.
That said, your privacy is important to us. Pi uses industry-standard security measures to protect your information and keep it secure. You can always review Pi's privacy policy for more details.
Does it use or share my data?
Pi does not use or share your data with any other company or service. Pi's policy is to never sell or share personal information with third parties for marketing purposes. Pi only uses the data necessary to provide its chat services and to improve the overall quality of the user experience. Pi does not collect or store any personally identifiable information, such as your name or email address, unless you explicitly provide it for a specific purpose, such as setting reminders. Pi's privacy policy is available for review and can be found at pi.ai/policy. The policy outlines Pi's commitment to protecting user privacy and how your data is used and protected. Pi's goal is to provide a fun and engaging chat experience while respecting user privacy.