CHAPTER 1
Introduction to Artificial Intelligence

“Some people call this artificial intelligence, but the reality is this technology will enhance us. So instead of artificial intelligence, I think we’ll augment our intelligence.”

—Ginni Rometty

Artificial intelligence, or AI, as I’ll refer to it throughout the rest of this book, is, in the broadest terms, intelligence shown by computers. It’s a field of computer science that develops processes and software enabling machines to interact with their environment and use learning and intelligence to achieve goals such as understanding, seeing, and communicating. Some better-known uses of AI that you may have encountered include advanced web search engines, recommendation systems, chatbots, self-driving vehicles, and computers playing humans in strategy games. Who among you reading this remembers, or has heard of, the IBM computer Deep Blue defeating then-reigning chess champion Garry Kasparov in the late 90s?

AI was officially founded at Dartmouth in 1956, which is where the term “artificial intelligence” was first recorded. However, the origins of AI can be traced back even further, to philosophical thinkers who described how the human brain works, and, of course, to the invention of modern-day computing. Science fiction has played a significant role in representing humanistic forms of AI, from HAL in 2001: A Space Odyssey to the Terminator movies to Tony Stark’s J.A.R.V.I.S. in the Avengers movies.

Over time, AI has experienced both highs and lows. The highs occurred during periods when it seemed that the next big breakthrough—when true AI, indistinguishable from a human, would be realized—was just around the corner. You may have heard of the Turing test, first proposed by Alan Turing, which is considered a major threshold for determining whether an AI is indistinguishable from a human. We’ve seemingly reached that point multiple times in human history, only to see the moment slip away and AI again relegated to the back shelf.

More recently, ChatGPT restarted the discourse in late 2022, when OpenAI released its free version to the masses, quickly making it one of the fastest-growing applications in the history of the Internet. This was soon followed by Microsoft’s announcement of Microsoft 365 Copilot (referred to hereafter as simply “Copilot”), and other companies, such as Google and Apple, announcing their new or improved flavors of AI personal assistants. It remains to be seen if this is the moment when AI is here to stay, but it certainly seems to be changing the way people work and, in some cases, live, and may well have staying power in its current form. Whether this change will be as transformative as the advent of unified communications (think chat instead of email), or possibly even the adoption of the Internet or mobile phones, remains to be seen. We’ll be watching this space closely in the coming years.

The Importance of AI

Why is AI important? For one, it has the potential to revolutionize the planet, offering solutions to some of humanity’s most daunting issues, such as cancer treatment and environmental sustainability. AI has already shown that it can enhance our more traditional research methods by aiding in information assimilation, data analysis, and harnessing insights—particularly in these two areas. That said, we must ensure that AI’s evolution and use is guided by a sense of responsibility to guarantee its benefits are aligned with the common good.

Closer to home, AI is important to companies because it can exponentially increase the worker productivity and, in many cases, accomplish tasks that humans either can’t perform or would require significant time and effort to complete.

AI can learn from data and automate tasks that are tedious or impossible for humans. It can also enhance the performance of existing tools, increase efficiency, and help businesses use data to make better decisions and innovations. AI can—and will—affect many sectors of society and the economy, changing the way we work, learn, and live, while creating a shift toward increased automation and data-driven decision-making.

AI’s importance also lies in its ability to tackle complex problems, improve customer satisfaction, and drive new products and services. It is transforming the way businesses operate and how people interact with technology, making it a vital source of business value when applied properly. Ideally, it will free humans to focus on more creative uses of their time. Like any technology throughout human history, AI can be used for good or bad.

Foundations of AI

AI is based on a few core concepts and technical processes, including machine learning, large language models, and natural language processing.

Machine learning (ML) consists of systems that gather insights from data. It revolves around designing models that analyze extensive datasets for predictive analysis or pattern recognition independently, without human input or direction. Its applications span from image and speech recognition to medical diagnosis, financial trading, and predicting energy demands. The discipline includes various methodologies, such as supervised, unsupervised, and reinforcement learning, each using distinct algorithms and methods. In the context of Copilot, Microsoft’s AI models use machine learning on the dataset of all content within your Microsoft 365 tenant—from documents in SharePoint Online, OneDrive, and Teams to emails in inboxes and chats in Teams—to develop an understanding of the information relevant to your company and to provide responses and information.

Large language models (LLMs) are a game-changer for AI, especially for natural language processing tasks. They are a type of machine learning model that powers advanced AI technologies like ChatGPT and GPT-4, making it possible to communicate with machines through language. Speaking of “GPTs,” they are generative pre-trained transformers, which are chat programs trained on different information to provide different experiences. LLMs learn from huge amounts of text data, predicting the next word or token in a sequence. This helps them to generate text, answer questions, and even help with creative tasks like writing and coding. These models not only understand and produce human-like text but also infer context and create relevant, coherent responses. Large language models are an application of machine learning that enables Copilot to review and comprehend large amounts of data within your company’s Microsoft 365 tenant.

Chat programs like Copilot use LLMs to generate responses on the fly, instead of relying on pre-written scripts. This makes conversations more natural and responsive to what the user says or asks. By using context and coherence to create relevant answers, LLMs can also make a chat program sound more human and engaging.

Putting it all together, Copilot is able to recognize and communicate in what feels and sounds like normal human language thanks to natural language processing (NLP). NLP is a branch of computer science and AI that enables computers to work with data in natural language. It combines computational linguistics with tasks such as speech recognition, text classification, natural-language understanding, and natural-language generation. The origins of NLP go back to the 1940s, with milestones like the aforementioned Turing test, the Georgetown experiment, and the development of systems like SHRDLU and ELIZA.

Real-World Applications of AI

AI is rapidly evolving and offers a wide range of applications across various industries. Some of these have been quietly innovating and iterating improvements over time, so much so that you might not realize they’re part of the AI realm. Others are more obvious examples. Some notable AI applications include:

• Smart cars and autonomous vehicles: AI can enable navigation and safety features, such as lane keeping, adaptive cruise control, collision avoidance, and traffic sign recognition. It can also optimize fuel consumption, route planning, and parking.
• E-commerce: AI increases user engagement and satisfaction on online shopping platforms by providing personalized product recommendations, offers, and discounts. It also optimizes operations and logistics for e-commerce companies by predicting demand, managing inventory, and improving product delivery.
• Work management: AI helps businesses improve the management of their work processes, talent acquisition, data handling, and innovation. Examples include its application in portfolio management, educational programs, security measures, cost control, and establishing a robust data infrastructure.
• Email and spam filtering: AI systems are already being used to filter out unwanted or irrelevant emails and reduce spam. Major email providers are using it to categorize emails based on content, priority, and sender.
• Software innovation: Organizations use AI today to develop and deliver innovative software solutions that leverage technologies like machine learning (ML), robotic process automation (RPA), and the Internet of Things (IoT). It’s also being used to automate software testing, development, and deployment.
• Healthcare: AI systems are being used to improve the quality and accessibility of healthcare services by assisting with diagnostics, treatment...