Understanding AI through national flags

Excerpt from the Book

Preface

On 15 February 2012, on my way to the UN Palais in Geneva to deliver a lecture on digital diplomacy, I checked my messages. To my surprise, artificial intelligence (AI) was on the students’ wish list of topics to be discussed. I was excited that somebody shared my interest in AI, a rather obscure topic at the time but one that had haunted me since 1992 when I wrote my thesis on expert systems for international law.  I was searching for a simple way to explain AI.

As I walked between the lanes of flags in front of the UN in Geneva, I had an Eureka moment: Why not explain AI through the patterns on national flags? 

The early origins of the flag explanation date back to lazy afternoons during summer holidays with my family. When the sun was too strong to enjoy water sports, and our stomachs were full after lunch, my daughter Zoe and I shortened the long Mediterranean siestas by playing a flag game: guessing the country’s name by seeing the flag. In this memory exercise, Zoe used a common pattern among flags as mnemonic shortcuts by grouping African, Arab, Nordic, Slavic, and other flag families.

During my Geneva lecture, students enjoyed my explanation of AI using national flags, and this book’s first lines were written. Since then, this story has evolved through hundreds of lectures and discussions. 

On 30 November 2022, the world discovered Chat GPT. AI became the talk of the day. The need to explain AI sky-rocketed.

As ChatGPT is based on a large language model (LLM), I extended my flag explanation to words. Thus, with numbers, a language of computers, flags and words became the pillars of this publication. 

I’m grateful to my daughter for the early inspiration to link AI and flags via patterns. I’m also grateful to my colleagues and students who helped with their comments and questions to sharpen these simple explanations of AI.

I am grateful to you, reader, for your readiness to dive into AI. The deeper we get into the conceptual origins of AI, the more we will walk the paths of centuries-old thinkers. We will realise that AI is as old as our eagerness to know, understand, and explain nature and society. As is written in the Book of Ecclesiastes, there is no new thing under the sun … including AI! 

Introduction

Understanding how AI functions is becoming necessary for everyone, from citizens to societal leaders. Citizens’ basic rights and well-being can be affected by AI. Political and business leaders have to make decisions on AI, sometimes with delicate trade-offs. AI poses two levels of explainability challenges. One is for technical people to know how neural networks function. The other is for the wider population to understand AI and what it is capable of. 

Why understanding AI functionality is important

As AI systems become more integrated into our everyday lives and critical infrastructure, we need to know how they function for several reasons:

Trust

AI technology has raised many fears worldwide, especially after the arrival of LLMs like GPT. For the sustainable development of AI systems and their full usability, people need to them. This trust will be nurtured with more understanding of how AI works and understanding the logic behind AI’s conclusions.

Cross-disciplinary cooperation

Future AI developments will depend on full cooperation among tech people, anthropologists, legal specialists, and other professionals. For this smooth, cross-disciplinary cooperation, it is critical to have a full and shared understanding of AI functionality.

Security 

AI triggers many security risks. Opaque AI systems increase security risks. Thus, understanding AI models through open-source and similar approaches can help identify bugs and potential misuse of AI-supported platforms.

Governance and compliance

AI as a black box cannot sustain pressure from governance and compliance. Regulators need to know how AI functions to ensure compliance and accountability. The EU’s AI Act and other forthcoming regulations will increase pressure for explainability. 

How AI is explained in this book

The explanation of AI is centred on three main protagonists: flags, words, and numbers. Flags are the main character. Words and supporting characters help us to understand ChatGPT and other AI generative systems that answer our questions and draft our letters. Numbers are the third character in this story. Two numbers – ‘1’ and ‘0’ – represent all we do with computers, including AI, through binary codes. Building on simple 1-0 (or yes/no), AI also uses more complex numbers and computations.

Our three characters use two major tools – patterns and probability – to generate AI answers. We can find patterns everywhere around us in nature, in towns, in our social relations, in our governance. In our story, we find patterns in flags, such as crosses in Nordic flags. We find patterns in words, such as the definition of diplomacy. Numbers also carry a lot of patterns from the Fibonacci sequence to modern statistics.

Once we identify patterns, we move to probability, the second main AI feature. Probability helps us guess what the next word would be in answering our question in ChatGPT. AI is one big guessing machine based on statistics and probability.

Following our three characters (flags, words, numbers) and two methods (patterns and probability) on this learning journey, we will find simple explanations for AI technology and terminology. We will borrow Occam’s Razor from philosophical enquiry to find answers with the smallest and simplest explanations. This approach can be described as KISS (Keep it Simple, Stupid) in more modern lingo. Our plain-vanilla explanations of AI technologies are supplemented with additional technical explanations for gradually diving deeper into machine learning, neural networks, and key aspects of AI developments. 

We demystify new AI lingo from foundational models to machine learning and deep learning, to name a few terms. 

We also move from the main road to the pathways of AI’s philosophical and mathematical underpinnings.

In the final section, we explain AI prompts as a practical way to interact with AI platforms.

Let’s start…