1.2 | Get to know AI and how it works

1. Introduction

In the last lesson, you learned that AI isn’t new and that scientists have been investigating the concept of "making intelligent machines"¹ since the middle of the 19th century.

In this lesson, you’ll learn about where AI is today. Specifically: what AI does and how it works.

To see how much you learn as you progress through this section, here’s another quick reflection question.

2. So, AI, tell me what you do

Let’s return to the conversation you imagined starting in Lesson 1.1. You’ve been introduced to your new acquaintance; you know their name and you’ve found out about their background. Now, it’s time to ask that go-to and often dreaded question: “So, what do you do?”

What does AI do? It’s a good question, but there’s one problem: we haven’t really defined what AI actually ‘is’ yet, so it would make sense to start there.

We’re going to encourage you to jump into the conversational deep end and ask it for yourself.

Feeling shy? Not sure you want to be so direct? That’s ok. We’ve asked AI for you. Navigate to the expandable box below to view the result.

Before you tick ‘define AI’ off your learning list, note that defining AI is a challenge for both scientists and policymakers, and many different definitions exist.2

The definition generated by Apolitical’s search tool (shared in the collapsible box above) is close to the Organisation for Economic Co-operation and Development AI Policy Observatory’s (OECD.AI) definition of an AI system.3

To help remember the technical terms you’ll come across, take a moment to download the AI glossary we’ve made for you.

3. The types of AI and how they work

Now, let’s take that definition we’ve given you, divide it by two, subtract one, go back a step and then take our findings in a completely different direction. Sounds complicated, doesn’t it? But we’re going to make it really easy for you to understand. You won’t need a calculator or an extra cup of coffee.

Helpfully, there are two main ways AI works:

1. Rules-based AI involves programming specific rules for each task it completes.

Remember ELIZA, the psychotherapist-chatbot from the previous lesson? ELIZA worked by taking the user's input and then searching for keywords or phrases that matched its rules. If a match was found, ELIZA would send a pre-programmed response. ELIZA was at the cutting edge of rules-based AI systems in 1964, but AI has come a long way since then.

1. Machine learning — unlike rules-based AI — doesn’t require computers to be pre-programmed.

Software like ChatGPT uses machine learning. For now, hold that thought, you’ll learn more about this in the next lesson.

Instead of being told explicitly what to do, machine learning enables computers to develop intelligence in a similar way to how humans develop intelligence: from being exposed to large amounts of data.

But before you get ready to present your handover notes, AI isn’t about to replace you or your colleagues. You’ll discover more about the differences between artificial and human intelligence in Section 3 of this course.

4. AI in your everyday and play

Now you know what AI is and an overview of how it works, it’s time to find out how it’s already being used in the world and how you’re already using it.

4.1 ‘Good Old-Fashioned AI’4

Starting life as ELIZA in 1964 and reaching its peak as Deep Blue (the computer that beat Garry Kasparov at a game of chess) in 1997, rules-based AI is now so integrated into everyday life that we no longer think of it as AI at all.5

Consider these everyday examples that use rules-based AI:

• Spam filtering uses a set of rules to identify emails that are likely to be spam by identifying keywords or suspicious email addresses.

• Self-checkouts use a set of rules to identify the products scanned by a customer and calculate the total cost.

• Traffic lights use a set of rules to determine when colours should change, such as when there’s a lot of traffic or when there’s an emergency vehicle approaching.

4.2 Learning machines

Today, ‘machine learning’ is what most people are talking about when they talk about ‘AI’.6 Machine learning is the type of AI we’ll focus on in this course.

Machine learning is relatively new, but it is becoming increasingly common in our everyday lives. It is used to power a wide variety of products and services you’ll be familiar with:

• Streaming services use machine learning to recommend movies and TV shows based on their user’s viewing history.

• Social media platforms use machine learning to personalise newsfeeds based on the user’s interests and activities.

• Virtual assistants like Siri and Alexa use machine learning to allow users to control devices with their voice.

Now that you’ve developed your understanding of AI, let’s revisit the first reflection question we asked you at the start of this lesson.

5. Conclusion

In this lesson, you’ve developed a basic understanding of what AI means today and discovered some real-world applications you’re already familiar with.

Now that you’re up to speed on the basics, let’s take a look at how AI is being used in the public sector.

Key takeaways:

• ‘AI’ is a broad term used to describe computer systems that can perform tasks that typically require human intelligence.

• AI systems are built using two main approaches: rules-based AI and machine learning.

• Rules-based AI is integrated into everyday life to the extent that most people no longer consider it to be AI.

• Today, ‘machine learning’ is what most people are talking about when they talk about ‘AI’.

1. McCarthy, John. “What Is AI?/Basic Questions.” Accessed August 11, 2023. http://jmc.stanford.edu/artificial-intelligence/what-is-ai/index.html.

2. Waller, Paul. “Defining ‘Artificial Intelligence’ for Regulation.” Apolitical, May 20, 2022. https://apolitical.co/solution-articles/en/defining-artificial-intelligence-for-regulation.

3. OECD.AI. “AI-Principles Overview.” Accessed September 6, 2023. https://oecd.ai/en/ai-principles.

4. This term was coined by John Haugeland in Artificial Intelligence: The Very Idea. MIT Press, 1989.

5. Luckin, Rose, Karine George, and Mutlu Cukurova. AI for School Teachers. CRC Press, 2022, p. 5.

6. Berryhill, Jamie, Kévin Kok Heang, Rob Clogher, and Keegan McBride. “Hello, World: Artificial Intelligence and Its Uses in the Public Sector.” OECD Working Papers on Public Governance. Organisation for Economic Co-Operation and Development (OECD), November 21, 2019. http://dx.doi.org/10.1787/726fd39d-en, p. 40.

⬅️ Previous lesson | Course outline | Next lesson ➡️

This is the logged out version of a short course in Apolitical’s Government AI Campus. To access all interactive elements and download a certificate of completion, sign up for free.