Topic Two (1.2): What is AI and How Can It Increase Productivity?

Objectives for this topic

By the end of this topic, you should be able to: 

  • Define Artificial Intelligence (AI) and its applications
  • Identify real-world examples of AI in various contexts.

You might also be able to: 

  • Reflect on ethical considerations related to AI
  • Explore the potential societal impacts and responsibilities associated with AI use.

Introduction

In this journey, you will explore the exciting world of AI and discover how it plays a crucial role in boosting productivity across various fields of expertise. Get ready to delve into the foundations of AI, explore its practical applications and contemplate the ethical considerations surrounding its use.

Mod 1 Topic 1.2 Pic 1.png

Working through your course

The video below outlines the activities that you are doing for this topic. (The video is aimed at the teacher but will give you an insight into what this topic is all about.)

You can also watch the video (below) discussed in the video above:

Now let’s look at the two types of programming used in producing AI mentioned in the film. Rule-based intelligence and data-driven intelligence represent two different approaches to achieving intelligent behaviour in systems, particularly in the context of AI and decision-making. Let's explore the key differences between these two types of intelligence:

Rule-Based Intelligence.

Definition:

Rule-based intelligence uses a specific set of rules and logical thinking to make decisions or solve problems. These rules are usually made by experts and contain clear information about the subject area.

Note. You could rewatch the first video above to help you with these characteristics.

Characteristics:

  • Explicit Rules: The system operates based on explicit rules that are defined in advance. These rules are often in the form of "if-then" statements, specifying conditions and corresponding actions.
  • Transparency: The decision-making process is transparent and understandable because the rules are explicit. Users can trace how the system arrives at a particular decision by examining the rules.
  • Limited Adaptability: Rule-based systems may struggle with adapting to new or unknown situations, since their behaviour is constrained by the predefined rules. If a scenario (situation) is not covered by the rules, the system may not provide a meaningful response.
  • Common in Expert Systems: Rule-based approaches are commonly used in expert systems where human expertise is codified (turned) into a set of rules to mimic the decision-making process of a human expert in a specific field.

Data-Driven Intelligence.

Definition:

Data-driven intelligence, also known as machine learning or statistical learning, relies on patterns and insights derived from data. Instead of explicit rules, the system learns from historical data and generalises patterns to make predictions or decisions.

Note. You could rewatch the first video above to help you with these characteristics.

Characteristics:

  • Learning From Data: The system learns patterns, relationships and trends from large datasets. This learning process allows the system to generalise and make predictions on new, unseen data.
  • Adaptability: Data-driven systems are more adaptable to changes and variations in the input data. They can potentially handle complex scenarios that were not explicitly covered during training, provided that similar patterns exist in the data.
  • Complex Decision-Making: Data-driven intelligence excels in complex decision-making tasks, especially when dealing with large and diverse datasets. It can identify hidden patterns that may not be apparent through rule-based approaches.
  • Black-Box Nature: The decision-making process may be less transparent compared to rule-based systems. The model learns complex relationships from data and the decision rationale might not be easily explainable in human-understandable terms.

Comparison of Rule-Based Intelligence versus Data-Driven Intelligence: 

  • Flexibility: Rule-based systems are less flexible and may struggle with novel situations, while data-driven systems are adaptable and can handle unforeseen scenarios based on patterns learned from data.
  • Interpretability: Rule-based systems are transparent and easy to interpret, while data-driven systems, especially deep-learning models, are often considered "black boxes" due to their complex internal representations.
  • Expertise Dependence: Rule-based systems depend heavily on human expertise for rule creation, while data-driven systems can leverage large datasets to automatically learn patterns without explicit human intervention.

In practice, a combination of these approaches is often used to harness the strengths of both rule-based and data-driven intelligence, creating more robust and versatile intelligent systems.

Example of a Rule-Based System:

Mod 1 Topic 1.2 Pic 2.png

Because Tic-Tac-Toe (noughts and crosses) only has a limited number of moves, we can create a rule-based intelligence algorithm to play the game, as in the picture above.

Chess, on the other hand, has more possible variations of games than there are atoms in the observable universe - about 1040

So, a deep-learning data-driven AI model has to be used.

Generally speaking, data-driven models recognise patterns in vast amounts of data in order to make reliable and statistical predictions.

Let’s look at some applications of AI systems:

Generative AI 

Generative AI is the term used to describe artificial intelligence applications that can generate content such as images, sound or text. Your task is to use an online application that uses AI to generate artwork based on the words or prompts that you give it.  

Instructions:

    • A poster for the world climate change conference 
    • The wall of a café or restaurant

Mod 1 Topic 1.2 Pic 3.png

While using the application:

  • Try generating the image again with the same prompts and the same art style. Does it generate the same artwork as last time?  
  • Try changing your prompts. Does it work better if more or fewer words are used? 
  • Think about the ethical questions behind you using this application. Does the art it creates belong to you? What art did it use to create this art? 

Computer Vision 

Computer vision is a field of AI that attempts to gain meaningful information from images. Your task is to use an online application that uses computer vision to attempt to identify the content of an image.   

The website below will allow you to choose an image. It will use AI to predict what is in the image.   

Instructions: 

Mod 1 Topic 1.2 Pic 4.png  Mod 1 Topic 1.2 Pic 5.png

While using the application: 

  • Why do you think there is a confidence rating? Is that important?  
  • Why do you think the application is more confident about some elements of images than others?

Review

Reflect upon what you have just covered. In your notes, summarise: 

  • What you have learnt 
  • What you already knew 
  • What surprised you 
  • What you are curious to know more about.

 Support activity for this topic

More on the difference between rule-based and data-driven (generative) chatbots: 

Extension activity for this topic

Follow the tutorial below to make your own UNBEATABLE rule-based Tic-Tac-Toe game in Scratch! Why not get one of your friends and family to play it! 

COMPLETE - Unbeatable Tic-Tac-Toe (instructables.com) Links to an external site.