Getting Started with AI: Basics and Beyond

Welcome to the "Getting Started with AI" webinar! Today, Ernest and I will guide you through the basics of artificial intelligence (AI) and beyond. We are from the Digital Futures Institute (DFI) at Teachers College, Columbia University. Teachers College (TC) is a Graduate School of Education, Health, and Psychology. I'm Debbie Bodry, the Director of Academic Technology, and my colleague is Ernest, the Senior Educational Technologist.

Agenda

Introduction to AI: What is artificial intelligence?
Demonstrations: A few demonstrations of AI tools.
Benefits and Challenges: Potential benefits and challenges of AI.
Beyond Basics: AI tools beyond text and images.

Disclaimer

The tools demonstrated have not undergone the TC review process, which checks for security, accessibility, and privacy. They are very new and often don’t offer the types of contracts that TC requires.

What Is Artificial Intelligence?

AI makes predictions based on data. You've likely been using AI for years in various forms:

Predictive Text: Suggestions while texting or emailing.
Voice Assistants: Siri, Alexa, etc.
Navigation: Google Maps.
Recommendations: Shopping sites, streaming services.
Social Media: Content based on your interactions.

Generative AI creates something new based on user prompts and training data. The most common text-based AI tool is ChatGPT, but other tools like Bing AI, Gemini, Claude, and more exist.

Let’s take a look at some examples.

Tour of ChatGPT

ChatGPT allows users to type prompts and get responses. For example, if you type "How do electric cars work?" it provides a comprehensive and quick response.

Large Language Models (LLMs)

LLMs learn from a vast amount of data on the internet. This training data enables the AI to generate natural-sounding responses.

Benefits and Challenges

Benefits

Writing: Idea generation, drafting, and revision.
Tutoring: Summarizing articles, explaining concepts, practice quizzes.
Educators: Lesson plans, text-to-speech.
Businesses: Shifting from lower to higher-level tasks.

Challenges

Inaccuracy: Sometimes provides incorrect information (hallucinations).
Bias: May perpetuate biases present in its training data.
Ethical Issues: Data sourcing, ownership, exploitation of resources.

Beyond Text and Images

Generative AI extends beyond text and images.

Video Generation

Runway can create videos from text descriptions.

Voice Cloning

Descript allows voice cloning, making a synthetic version of your voice.

Digital Avatars

Synthesia and HeyGen create digital avatars based on short video recordings.

Examples

Runway

Creates a video from a text prompt like "create a video of two-year-olds playing in a classroom."

Descript

Allows you to generate synthetic voices by training on a short recording of your voice.

Synthesia

From a short webcam recording, creates a digital avatar that can read any given script.

AI in Mainstream Products

AI tools are being built into widely used products:

Microsoft Co-Pilot: AI assistant for Microsoft Office.
Google Workspace: Gemini for Google Docs and Sheets.
Padlet: AI-generated images for digital post-it notes.

Where to Begin

Free tools like ChatGPT 4.0 and Adobe Firefly are excellent starting points but avoid inputting confidential information. Always be aware of the limitations and biases.

Upcoming DFI Events

Next week, we will focus more on image generators with artificial intelligence, and our fall events will be posted soon.

Thank you for attending!

Keywords

Artificial Intelligence
Generative AI
Large Language Models
Predictive Text
Voice Assistants
AI Writing Tools
Digital Avatars
Video Generation
Voice Cloning
Ethical Issues

FAQ

What is Generative AI? Generative AI creates new content based on user prompts and its training data.
Which AI tools were demonstrated? Tools like ChatGPT, Adobe Firefly, Runway, Descript, Synthesia, and HeyGen.
What are Large Language Models (LLMs)? LLMs are AI systems trained on a vast amount of text data from the internet, allowing them to generate human-like text.
What are some challenges associated with AI? Challenges include inaccuracy, biases, ethical issues regarding data sourcing, ownership, and resource exploitation.
Where can I start with AI tools? Free tools like ChatGPT 4.0 and Adobe Firefly are good starting points, but avoid inputting confidential information.
How is AI being integrated into mainstream products? AI is being integrated into products like Microsoft Office and Google Workspace to assist with writing, analyzing data, and more.

By following these steps, you can better understand the basics and beyond of AI, use it responsibly, and navigate the opportunities and challenges it presents.