1.1 Overview of AI and Machine Learning — The AI-Enhanced Classroom: Redefining Teaching and Learning

Guiding Questions:

What are the historical milestones of artificial intelligence ?
What are basic concepts and theories in artificial learning?
What is generative AI and how does that impact education?

Let’s Get Started!

Welcome to Pedagog.ai’s comprehensive course, The AI-Enhanced Classroom. We are thrilled to have you join us in exploring the exciting world of artificial intelligence and its potential applications in education. In this first module, we will provide you with an overview of artificial intelligence, its historical milestones, various types of AI technologies, and recent advancements in the field. We hope to spark your curiosity and encourage you to consider how these technologies might impact your teaching practice and enhance your educational experiences.

A Brief History of AI

Artificial Intelligence (AI) has been a part of our society for many years, dating back to the 1950s when the first computers were being developed for problem-solving purposes. Some notable milestones in the development of AI include:

1950: Alan Turing proposes the Turing Test, a method for determining whether a machine can exhibit intelligent behavior.
1956: The Dartmouth Conference, where the concept of artificial intelligence is formally introduced.
1959: The term "machine learning" is coined by Arthur Samuel.
1970s: The development of the first expert systems, computer programs capable of making decisions based on previous experiences.
1980s: The introduction of neural networks and natural language processing (NLP), enabling machines to better understand and process human language.
1990s: Integration of AI with other technologies such as robotics, leading to the development of the first driverless cars.
2000s: AI advancements in facial recognition, medical diagnosis, and many other applications.
Present: AI technologies like Chat GPT, a generative AI, continue to shape the skills needed for success in education and the workplace.

Understanding this historical context is essential for appreciating the potential of AI and its applications in various fields, including education.

Artificial Intelligence

Artificial Intelligence (AI) is an interdisciplinary field that seeks to create machines capable of mimicking human-like cognitive abilities such as learning, reasoning, problem-solving, and decision-making.

There are different categories of AI depending on what exactly its capabilities are.

The first kind, referred to as Narrow AI or Weak AI, is quite common. It focuses on performing specific tasks, like recognizing your voice when you chat with Siri. This type of AI has a limited context, meaning it doesn't have a broader understanding or awareness. You may also encounter AI that is more specialized in certain fields, such as diagnosing diseases or translating text.

Currently, Narrow AI is the most advanced form of AI we have, but generative AI systems like Google’s Bard and OpenAI’s ChatGPT have sparked debates about our proximity to achieving Artificial General Intelligence or AGI.

AGI, sometimes called Strong AI, refers to AI that can perform any intellectual task that a human can. It can understand, learn, adapt, and apply knowledge in a broad range of tasks, much like a human being.

Then there's Superintelligent AI, which goes beyond AGI and is, theoretically, smarter than humans in all aspects. This includes creativity, wisdom, social skills, and understanding human thoughts and emotions. Superintelligent AI, the topic of sci-fi movies, would be conscious and self-aware, much like humans. This level of AI doesn't exist yet, and its potential realization raises many ethical questions.

So, how does AI come to life? The building blocks of AI are algorithms, which are trained using data to improve their performance—this process is called machine learning. In other words, AI learns from examples in the training dataset, identifying patterns and relationships. This learning forms the basis for AI to make predictions or decisions. You might have heard about different AI models like GPT-4 and Bard—think of these as turbo-charged predictive tools, much like your phone's autocomplete!

The selection and quality of the training data are crucial, as any biases or gaps in the data can surface in the AI models. The models are then deployed to make predictions or decisions on new data. This is the point at which most of us interact with AI, even if we aren't computer scientists.

AI can learn in several ways:

Supervised learning: Here, the AI is fed labeled data to learn from and make predictions. Imagine teaching it to recognize animals in pictures by showing it thousands of tagged images.
Unsupervised learning: This is when AI finds patterns in unlabeled data. It's like giving it a library of books to group based on subject, reading level, and language without any guidance.
Reinforcement learning: This is when an AI learns by doing. It's like an AI learning to win a game by playing it repeatedly and adjusting its strategy based on the results.
Deep learning: This is an advanced form of machine learning inspired by the human brain. The AI processes information through multiple layers—through multiple scaffolded steps—modifying the information based on rules it learns from the data.
Transfer learning: This involves using a pre-trained model and adapting it for a new but related task. It's like training an AI on a large image classification task, then tweaking it to recognize specific objects.

Understanding the creation of AI models can shed light on their potential weaknesses and opportunities, especially in education. As we delve deeper, you'll start to see the direct connection between these principles and the application of AI in the world of education. Understanding the fundamentals of AI will also help you talk about AI with your students whether it be in the context of teaching AI literacy or helping students themselves understand the capabilities and limitations of the tools they are using.

Generative AI

Generative AI refers to a class of AI models that can generate new content, such as text, images, or music, by learning patterns and structures from existing data. For educators, it's important to have a general understanding of generative AI and its potential applications in the classroom. Here, we provide a broader, less technical perspective on generative AI.

Generative AI stands out as one of the most intriguing and promising realms in the wide spectrum of artificial intelligence. Its primary function revolves around learning from vast amounts of data, such as blocks of text, collections of images, or any type of digital information. During its training, it identifies intricate patterns or structures within this data, forming an understanding of how the components of the data interact and relate with each other.

The power of generative AI shines when it begins to create or 'generate' new content. Unlike simple replication or copying, generative AI uses its understanding of the underlying patterns in the data to create entirely new content that bears striking similarity to the original data but is uniquely different in its specifics. Whether it's crafting convincing text, creating visually appealing images, or even composing music, these models produce content that is not just coherent but often surpasses the threshold of what we perceive as 'realistic'.

However, to truly appreciate the capabilities of generative AI, it is essential to understand a couple of foundational principles:

Representation Learning: This refers to the AI's ability to automatically find and learn the significant features or representations from raw data needed to perform tasks. This skill is critical for the AI to make meaningful interpretations and predictions from the data.
Generative Modeling: This concept is about the AI's ability to generate new data instances similar to those in its training set. Generative models do this by capturing the probability distribution of the training set. In simpler terms, it tries to understand how the data in the training set is spread out and what features are common or rare. Armed with this understanding, it can then generate new data that fits this distribution.

With these capabilities, generative AI models not only analyze and learn from data but can also imagine and create. They can turn a few brush strokes into a painting or a simple melody into a symphony. They can generate realistic dialogue for a novel, or propose multiple design concepts for a new product.

However, because these generative AI models cannot “think,” they can also sometimes make things up, a phenomena called “hallucination.” In the context of generative AI, "hallucinations" refer to the instances where the AI generates outputs that are not strictly accurate or factual, but are fabricated based on the patterns and structures it has learned from its training data. These inaccuracies underscore the importance of using high-quality training data and carefully tuning the AI model to ensure reliable and useful results.

Generative AI models have numerous potential applications in education, including:

Content generation: Generative AI models can create realistic text, images, or other content that can be used for lesson plans, quizzes, or assignments. This can save teachers time and effort in preparing materials and provide students with diverse learning resources.
Personalized learning: Generative AI models can analyze students' learning patterns, strengths, and weaknesses to create personalized learning materials and activities tailored to each student's needs.
Creative exploration: Students can use generative AI tools to explore their creativity in various domains, such as writing, art, music, or design. By working with AI-generated content, students can develop new ideas and perspectives, enhance their creative skills, and gain a deeper understanding of the AI technologies shaping their world.

Generative AI also has implications beyond their use in the classroom. As society continues to integrate generative AI tools, the kinds of skills and knowledge that students need in the real world will shift.

Lesson 2 covers more about both the potential applications in education and the implications on society.

For now, here’s an example of what interacting with a generative AI tool looks like:

https___chat.openai.com - 31 May 2023

Key Takeaways

The development of AI and machine learning technologies has spanned several decades, with numerous historical milestones that have shaped the field.
AI technologies, such as expert systems, neural networks, natural language processing, and generative AI, have the potential to transform various aspects of education and the workplace.
Understanding the foundations and different types of AI technologies is essential for educators who wish to integrate these powerful tools into their teaching practices and adapt to the rapidly evolving technological landscape.
Generative AI models learn patterns and structures from data and can generate new, realistic content that resembles the original data.
Teachers should be aware of generative AI's potential applications in education, such as content generation, personalized learning, and creative exploration.