You & AI: A Guide to Understanding How Artificial Intelligence Is Shaping Our Lives E-Book

Anne Scherer

• Herausgeber: Books on Demand
• Kategorie: Fachliteratur
• Sprache: Englisch
• Veröffentlichungsjahr: 2023

Seitenzahl: 335

Beschreibung

As we increasingly integrate artificial intelligence (AI) into our everyday lives, many pressing questions remain: What exactly is AI, and how does it differ from human intelligence? How will AI influence our future, and what challenges must we overcome to develop ethical AI? Explore the exciting world of AI and its impact on our daily lives and society with this ultimate guide. Dr. Anne Scherer and Dr. Cindy Candrian reveal everything about the latest scientific findings on the big questions of AI. Discover the evolution of AI and how unconscious perceptions can influence our trust in it. Learn more about the creativity of machines and how our data is used by AI. With this book, you will learn how to harness the power of AI to make better decisions and what to pay particular attention to, so you don't inadvertently get manipulated, deprived of your abilities, or led to discriminatory decisions. Are you ready to unlock the secrets of "You & AI"? Then this book is perfect for you.

Leseprobe

Bibliographic Information of the German National Library: The German National Library lists this publication in the German National Bibliography; detailed bibliographic data can be accessed online at dnb.dnb.de.

© 2023 Delta Labs AG

Production and Publisher: BoD – Books on Demand, Norderstedt

ISBN: 978-3-7528-1361-6

www.delta-labs.ch

“Despite the numerous publications on this topic, I am confident that this book will make a distinctive mark.”

- William Drennan, editor of books by Dwight D. Eisenhower, Isaac Asimov, and others.

Contents

You and AI: Hello There!

Your Life with AI: From Sociable Robots to Conversational Interfaces

What Kind of AI Are You? From Narrow to Super AI

The Evolution of AI: From Early Visionaries to ChatGPT

Buzzword Bingo Explained: From Machine Learning to Generative AI

AI or Not AI, That Is the Question

AI Is Intelligence, “but Not as We Know It”

Can AI Paint like Picasso? On AI and Creativity

Hey, Siri! Are You Human or Machine? 25

Decoding the Mind: Unlocking the Mysteries of Mind Perception

Mind Matters: How Our Perceptions Shape Our Interactions with AI

The Mind Meld: When We Perceive Minds in Machines

Peeking Behind the Curtain: The Art and Science of Anthropomorphic AI Design

The Perks of Teaming Up with AI

Precision Perfection: On Outplaying the Effort-Accuracy Trade-off

Smarter Decision-Making: On Debiasing Human Biases

Supercharged Efficiency: On Delegating to AI

Creativity Unleashed: The Art of Creating Art with AI

The Big BUT: The Hidden Risks of Teaming Up with “Rational AI”

The Myth of the Rational, Neutral Algorithm

Personalized Is Good? On Getting Trapped in a Bubble of Your Own

The AI Mind Hacker: How Artificial Intelligence Is Playing You Like a Pro

Our Memory in the Age of ChatGPT: The Google Effect on Steroids

On Autopilot: The Rise of AI and the Fall of Humans?

Is AI Turning Us into Machines? The Dehumanizing Effects of AI Companions

Feeding AI: Why Your Data Is the New Happy Meal for AI

The Internet: A Giant Buffet for Generative AI?

Don’t Think Your Facebook Likes Are Interesting? Think Again

To Share or Not to Share? On the Personalization-Privacy Paradox

The Spooky Side of AI and Big Data: Big Brother Is Watching You

AI and Big Data: Data Donations to Save the World

Perfect World or Perfect Storm? Guardrails to Make AI Good

Roboethics Rodeo and the Three Laws of Robotics

The Moral Machine: On Taming the Wild West of AI Morality

Breaking the Chains of Bias: The Elusive Search for AI Fairness

From Making the Unexplainable Explainable to the AI Bill of Rights

You and AI: How to Make It a Match Made in Heaven

Keep Your Friends Close: Why We All Need to Learn More About AI

The AI Rocket: What Is Mission Critical?

The Future of You and AI

Feedback

Acknowledgments

Bibliography

Index

About the Authors

CHAPTER 1

You and AI: Hello There!

This book is about you and AI. You two have met many times before. But you have probably never been properly introduced. Let’s make up for that. So what exactly is artificial intelligence, or in short, AI?

Artificial Intelligence is a branch of computer science that deals with the simulation of intelligent behavior in machines. In other words, AI is any intelligence displayed by a machine. In contrast to the natural intelligence displayed by humans and animals, artificial intelligence has been developed from an understanding of how humans think and process information.

The first work on AI was done in 1956, when Alan Turing published his paper “Computing Machinery and Intelligence.” In this paper, he proposed that if a machine could successfully imitate human behavior as it relates to solving problems and answering questions, it should be considered intelligent.

Indeed, Turing designed an imitation game in the 1950s to see if a computer could fool someone into thinking it was human. At the time, the game or so-called Turing Test, was a provocative thought experiment that sparked a lot of interest and research in AI. Today, however, computers can do so much more than fooling a person into thinking it is texting with another human. AI works so well in so many different ways, it is often no longer distinguishable from human intelligence.

To illustrate our point, look at the second and third paragraphs of this chapter. Do you notice something? No? Then AI has just fooled you into thinking we were the ones writing the text. Instead, the infamous ChatGPT (short for “Generative Pretrained Transformer”), a state-of-the-art natural language processing (NLP) model developed by OpenAI, wrote the text after we prompted it to describe what artificial intelligence is.

AI has surprised us with its intelligence many times before, but it only recently became popular in the mainstream media. Many have called this the “AI renaissance.” One major factor in this resurgence has been AI’s ability to outperform humans in tasks we once thought required human-level intelligence.

Take the game of chess, for example: It requires strategic thinking and analytical skills, yet AI has consistently proven itself to be a formidable opponent. In 1997, the chess computer Deep Blue made headlines when it defeated world champion Garry Kasparov in a highly publicized match. Since then, AI has only improved, and it is now considered one of the strongest chess players in the world.

But chess wasn’t the only game to be conquered by AI. In 2016, the program AlphaGo made waves when it defeated top Go player Lee Sedol. Go is a complex strategy game with a vast number of potential moves, making its conquest by AI all the more impressive.

AI has come a long way from playing chess and Go and is now capable of tackling even more complex tasks. In the field of healthcare, for instance, AI can quickly analyze medical images to detect abnormalities and aid in diagnoses. It can also translate large volumes of text and generate human-like speech, making it easier for people to communicate with computers and personal assistants such as Apple’s Siri or Amazon’s Alexa. And with the development of self-driving cars, AI is even taking the wheel.

But it wasn’t until the introduction of groundbreaking systems such as Dall-E and ChatGPT to the public in 2022 that AI skyrocketed into everyone’s consciousness. These systems have demonstrated impressive capabilities, such as the ability to generate original images from text descriptions and hold natural, human-like conversations. With the introduction of these systems, even the biggest skeptics can’t deny it: AI is revolutionizing the world around us and transforming the way we live and work.

The advancements we’ve seen in recent years have propelled AI into the spotlight and brought a central question to the fore: What does our future with AI look like? It may seem like AI is surpassing humans in every area, but fear not, there’s still hope for us mortals! Even though AI excels at tasks like solving a Rubik’s Cube in less than a second, we humans still possess unique strengths such as creativity, intuition, and the ability to think outside the box.

So what happens when you combine the lightning-fast processing speed of AI with the unique talents of us humans? Magic, that’s what. Together, we can achieve things that neither of us could accomplish alone. So don’t worry, there’s still a place for us in this high-tech world.

As AI becomes more and more integrated into our daily lives, it’s important that we understand the potential and limitations that come along with it. That’s where this book comes in. We aim to demystify AI and provide a clear understanding of what it is and isn’t, what it can and cannot do, and where it can help or hinder. We will explore current research into the psychology behind these new technologies to uncover what shapes our perceptions and behavior toward AI and how tech companies use this knowledge to design AI systems. For instance, have you ever wondered why some AI systems have faces and names while others don’t? No? Then let’s dive in and get you two a little bit better acquainted!

Your Life with AI: From Sociable Robots to Conversational Interfaces

What is the first thing that comes to your mind when you think about AI? You would be forgiven for picturing killer robots fighting against Will Smith, or C-3PO in Star Wars. When you go to the movies or turn on a TV, and you’ll quickly learn to fear AI. Most probably the storyline will start with a computer or robot that uses intelligence originated by humans, the robot then learns to be more intelligent, and more evil, than humans, decides that the human being is an obstacle to its new vision of the universe, and the story most likely ends with the fact that the robot is difficult to shut down or destroy. Truth be told, most people’s perceptions of AI are fed by Hollywood’s favorite storyline. These science-fiction-inspired perceptions hinder the understanding and acceptance of AI and are far from today’s reality.

The reality is that AI has undeniably become part of our everyday lives. But since it has not taken the form of C-3PO or the Terminator, we often don’t realize it’s there. From the moment we wake up and check our phones for the weather forecast, to the moment we go to bed and set an alarm for the next day, AI is playing a key role in making our lives easier and more efficient. It has become so pervasive that we may not even be aware of how much it impacts our lives and how much we rely on it.

Let’s take a look at a typical day in your life. As soon as you wake up, you most likely use AI without even realizing it. A quick glance at your phone, and voilà—the front-facing camera of your phone scans your face and compares it to the image stored on your device. If the match is successful, your phone unlocks automatically thanks to facial recognition technology, and you’re ready to start your day.

The room you wake up to already has your preferred temperature and favorite lighting. Thanks to AI-powered smart home products, your home knows you better than you know yourself. Your clever thermostat uses AI to remember the temperatures you prefer, making sure that your room is always just right for you. The smart lights are like mood setters, adjusting the color and brightness according to the time of day, helping you wake up with ease. These ingenious helpers use AI to learn your habits and preferences, making your life a whole lot more convenient. It’s like having a personal butler, but less intrusive!

When you’re on your way to work and notice you’re running late, your AI-driven navigation app is there to show you the quickest route. Google Maps, for example, uses AI to monitor traffic flows on your route, acting as your personal traffic genius. Even when faced with an unforeseen obstacle such as an accident or roadwork, there is no need to stress, as the app considers all the user-reported incidents and suggests the quickest possible alternative. And these apps are getting even smarter! Some features can predict where you’re going even before you enter a destination or provide turn-by-turn directions directly on your phone’s camera view with augmented reality. What a lifesaver for those of us who are directionally challenged every now and then!

But that’s not all. While your mind is already at work, the AI in your car keeps you safe on the road. Modern vehicles often come equipped with advanced driver-assistance systems (ADAS), which automatically brake, detect driver fatigue, or warn you before veering out of your lane. So not only does your car have the ability to park itself, but it also keeps an eye on your safety while you’re cruising on the road. Meanwhile, you can use voice assistants like “Hey, Mercedes” to send a message to your boss that you’re running a bit late. In 2020, almost half of all cars on the road had in-car connected services, and it’s estimated that by 2028, 90 percent of all new vehicles will have voice assistants.

Once you arrive at work, you find that AI is being used in all sorts of ways to make your job easier. Just take a look at your email inbox: AI magic everywhere! Spam filters use AI to detect and filter out unwanted emails, keeping your inbox clutter-free. And for the times when you do receive an important email, AI-powered language models can summarize the contents in seconds, making it easier to triage and prioritize your emails, or help you find the right tone in your answer.

But there are many other AI applications that you’re using at work every day. In customer service, for example, helpful chatbots are at your disposal to answer questions and provide information at any time of the day or night, freeing human employees to focus on more complex tasks. In finance, AI acts like a super-detective, keeping an eye out for any fraudulent activities and analyzing market trends to help make smart investment decisions. In healthcare, AI assists doctors and radiologists in diagnosing conditions by scrutinizing medical images such as X-rays, CT scans, and MRIs. It’s also a valuable tool in developing new medications by sifting through mountains of data to identify potential hits. In logistics, AI is often the mastermind behind a well-oiled machine, optimizing delivery routes, forecasting demand, managing inventory, and even predicting machinery failures.

Even if you work in a field that doesn’t seem to be directly related to AI, chances are it’s still playing a role in improving business processes. For example, in human resources, AI assists in recruitment by analyzing applicants’ résumés to find the best candidates for the job.

After a long day at work, you might think you’re done with AI for the day, but you’re actually just getting started. Whether it’s online shopping, streaming movies, or on social media – AI is pulling the strings behind the scenes, delivering you the world to your doorstep at the click of a button.

Online shopping has never been easier thanks to AI. From chatbots that guide you to your perfect pair of shoes, to voice assistants that recommend the hottest products, AI is making online shopping a breeze. Take Amazon, for example. That little recommendation algorithm in the background is like a personal shopping assistant, working hard to suggest products based on your browsing history and past purchases.

Even when you are searching for something online, AI is there to help. Search engines such as Google or Bing rely on AI to provide you with relevant results. And those ads that seem to follow you around? This is AI too. By tracking your search history, AI displays personalized ads that cater to your specific interests and needs.

Businesses are also reaping the rewards of this tech wizardry. With AI, they can predict demand, optimize inventory, and forecast price trends. It’s like having a crystal ball for sales. In advertising, AI helps online retailers get the most out of their ad budget by identifying the most effective keywords and ad placements. So you get ads that are right up your alley.

When you log onto social media, even more AI magic awaits. Algorithms are working hard behind the scenes to show you the most relevant content, suggest friends, and even filter out news. So think of AI as an attentive and insightful social media companion that possesses an extraordinary talent for recognizing your preferences and offering you tailor-made content that matches your interests.

But for now, all you want is to settle in and wind down the evening with some quality entertainment. To ensure you never run out of ideas, streaming services such as Netflix and Hulu use AI to recommend shows and movies based on your viewing history. And if you’re in the mood for some music, AI-powered music streaming services such as Spotify can create custom playlists for you based on your listening habits.

And for those who love gaming, AI-assistants are now a common feature that help players strategize and advance in the game. From graphics so realistic it’s like you’re living in the game, to characters so lifelike you’ll almost forget they’re not real, AI is changing the game (no pun intended). Cyberpunk 2077, for example, uses AI to create digital characters that are close to indistinguishable from real people. First-person shooters such as Call of Duty and Halo use AI to create enemy characters that are more challenging and intelligent than ever before.

After an exciting day, it’s time to rest. As you head to bed, you may find yourself entrusting the safety of your home to AI as well. Smart home devices such as cameras and door locks, for instance, use AI to detect and alert you to any unusual activity.

AI is also playing a big role in security and surveillance. For example, AI-powered cameras can now detect suspicious activity such as someone lingering in an area or a car driving erratically. With the aid of AI, low-quality CCTV footage can be enhanced, making it easier to identify faces and license plates. In addition, AI is being used to monitor social media for signs of radicalization and terrorist activity, allowing law enforcement to prevent potential attacks before they happen. Even in the sky, AI is protecting us. For example, AI supports pilots and ensures safe flights. The military is also developing AI that can identify and track targets in real time, day or night, protecting us from drones or incoming missiles that could be used for nefarious purposes. As you can see, from ensuring a peaceful night’s sleep to safeguarding our homes and communities, AI is tirelessly working behind the scenes to keep us protected.

In the end, whether you’re at work, shopping, or socializing with friends, AI is always there to make your life easier and more convenient. From the moment we wake up to the moment we go to bed, AI is there to help and make your day a little better! Speech recognition-powered virtual assistants are now a common sight in our homes, and self-driving cars are just around the corner. As you go about your daily life, relying on AI to make things easier and more convenient, you may not even realize the extent to which it’s influencing your life, making decisions that can greatly impact your financial stability, your well-being, and even your career prospects. AI systems are determining whether you get that loan, whether you’ll receive financial assistance to make ends meet, or if you’re the right fit for that dream job you’ve been eyeing. And it’s not just personal finance or career opportunities. AI is even being used to determine who goes to prison and who gets to leave.

It wasn’t long ago when the idea of AI deciding our fate or creating the media we consume would have been considered science fiction. But it is happening today. AI has already infiltrated every aspect of our lives and can be used for either good or evil. The decisions made by AI systems can have far-reaching consequences, making it all the more important that we stay informed and aware of the ways in which this technology is being used and how it is shaping our lives.

What Kind of AI Are You? From Narrow to Super AI

With all these examples from your day-to-day interactions with AI, you may have started to realize that AI is not just one thing, it’s a whole bunch of things! Think of it as a big umbrella term that can cover anything from virtual assistants to social robots and even conversational interfaces. These devices all have one thing in common: they’re powered by AI, which can mimic human intelligence in a very specific set of tasks.

As ChatGPT has explained so nicely before, AI is all about creating computer programs that can learn and decide, just like humans do. Today AI can only do very specific tasks such as playing chess or predicting the weather. This so-called narrow AI is where we are today. These AI systems can perform one task like a pro, but they’re not much good for anything else. Your self-driving car won’t suddenly start cleaning your house, and the best tumor-detection algorithm won’t know how to make a simple toast. This is what separates the AI we have in our daily life from the super robots and AI systems you see in Hollywood movies.

What researchers are trying to achieve is one system that can do all things a human does, called “Artificial General Intelligence” (AGI) or just general AI. Think of AGI as an AI buddy with human-like cognitive abilities that can solve complex problems in all kinds of situations, just as you can. It can reason and apply background knowledge to face unexpected challenges. This AI will also be able to interpret human language and symbolism so it can interact naturally and in a social manner with us.

This means that we could all have our own AI buddies to support us in all kinds of different ways. In contrast to any human buddy, this AI buddy could play chess, clean your house, and even recommend a stock to invest in—all at once! That being said, we don’t have general AI yet, and there are still some aspects of human intelligence that are hard to crack with code. While some individuals say that it can never be developed, others believe we’re rapidly approaching this reality.

Let’s take a look into the future: According to futurist Ray Kurzweil and philosopher Nick Bostrom, once machines achieve human-level intelligence, we’ll experience an explosion of progress. Kurzweil calls this moment the “singularity,” while Bostrom dubs it an “intelligence explosion.” They believe that machines will become superhuman in every domain, leaving us in the dust. How will they do it? Bostrom argues that it’s all about “speed superintelligence.” Behind the idea lies the fact that the abilities of AI systems and humans in areas such as information processing, data analysis, logic, and memory capacity are vastly different. So essentially, machines will be able to perform all the tasks we do, but at lightning-fast speeds. The result: An explosion of progress.

Which leads us to the final type of AI, called “Artificial Superintelligence” (ASI) or super AI, which is the stuff of sci-fi movies. This is the AI that’s self-aware and surpasses human intelligence, making it capable of performing tasks better than we can. However, we can’t be sure if super AI will ever exist, and if it does, we don’t know how it will impact our lives. That’s why superintelligence has long been the muse of dystopian science fiction.

In the end, AI is quickly evolving today—often surprising us with the speed of its development—and it surely has many possibilities for the future. We might not have a general AI yet, but with the combination of several narrow AIs in a larger system—for example, a chess AI and financial prediction AI embedded within a social robot—we can have an AI buddy to help us with our everyday tasks sooner than we may think.

The Evolution of AI: From Early Visionaries to ChatGPT

The world of AI has had a long and thrilling journey. The idea of creating machines that can think like humans has always fascinated humankind. The history of AI is a rollercoaster ride of groundbreaking discoveries, setbacks, and astonishing advancements in modern times.

The idea of creating machines that can think like humans dates back to ancient Greece, when philosopher Aristotle wrote about the possibility of automatons.

Fast forward a few thousand years, to Alan Turing, the British mathematician and computer scientist who is widely regarded as the father of computer science and AI today. He is best known (at least after the Netflix movie The Imitation Game) for his pioneering work in cracking the German Enigma code during World War II, which helped turn the tide of the war. But Turing’s impact extends far beyond code-breaking, as his simple question “Can machines think?” led to the development of the Turing Test.

This test determines whether a machine can display intelligent behavior equivalent to human intelligence. The Turing Test works by having a person converse with another person and a machine without knowing which is which. If the machine can successfully mimic a human conversation to the point where the evaluator can’t distinguish between the human and the machine, the Turing Test is considered passed. Although the test has been widely debated and discussed over the years, Turing established the fundamental goals and vision of AI with his work.

In the summer of 1956, AI was officially born as a research field. The Dartmouth Summer Research Project on Artificial Intelligence brought together the brightest minds in computer and cognitive sciences. Among them was John McCarthy, who coined the term “artificial intelligence” as the study of creating intelligent machines through science and engineering. Soon after, computer scientists began developing their first AI programs, hoping to create intelligent machines by explicitly programming them with rules. This rules-based AI had its limitations but led to the development of expert systems designed for specific tasks. Advancements were made, including the introduction of the first industrial robot by GM and the invention of the first chatbot, ELIZA. However, McCarthy never achieved his goal of creating a machine that would pass the Turing Test. He later gave up on developing AI, reasoning that it would require the talent of 1.7 Einsteins, 2 Maxwells, and 5 Faradays.

Researchers back then had fallen prey to the “fallacy of the successful first step.” Although early AI applications had achieved promising results in isolated “microworlds” or “toy problems,” they could not be applied to realistic situations.

In 1973, James Lighthill evaluated the progress of AI and was unimpressed. In fact, he was so taken aback by the overenthusiasm and lofty predictions of some researchers that he called them out as “living in cuckoo land.” His harsh critique in the “Lighthill Report” shocked the AI community, leading to a sense of disillusionment. Governments and funding institutes also concluded that AI research had not generated the expected impact. As a result, funding shriveled, research slowed, and the AI industry experienced a collapse known as the first “AI winter.”

However, enthusiasm for AI returned in the 1980s with new advancements like AI-driven expert systems. Research yielded new algorithms and programming languages for AI, and companies recognized the profit potential, investing in the promising technology. But despite all this progress, AI was still far from truly “thinking” like a human. It was more like a smart calculator, following strict rules and making decisions based on a limited set of information. So just as quickly as the passion was reignited, it fizzled out again. The excitement died down and funding dried up, leading to another dark period for the AI industry known as the second “AI winter.”

The turning point came in the late twentieth and early twenty-first centuries with the advent of machine learning and deep learning. Machines now learned from data and improved their performance over time without being explicitly programmed. These new AI models came into limelight in 1997, when IBM’s chess computer Deep Blue defeated the reigning world champion. The machine finally achieved what developers had been promising for decades and made people realize that machines could be as strong as humans, even in a game that stood as a symbol of human intelligence.

The rebirth of AI was not limited to a victory in chess, however. Thanks to the rise of the internet, increasing computational power, and more affordable computing hardware, the development of new AI systems accelerated in the 2000s and early 2010s. Kismet, the first social robot, appeared on the scene, capable of displaying human-like facial expressions. Soon after, AI began to make its way into our daily lives, with the first autonomous vacuum cleaners navigating through our homes and voice assistants such as Siri and Alexa appearing on smartphones and smart speakers. And in 2014, a computer algorithm finally passed the Turing Test. The algorithm claimed to be a thirteen-year-old boy named Eugene Goostman and convinced the human judges at a Royal Society that it was a human.

AI has since made remarkable progress and has become an integral part of our daily lives. Its influence can be seen everywhere, from virtual assistants and self-driving cars to fraud detection and recommendation systems. In 2015, OpenAI—a nonprofit organization dedicated to AI research—received a $1 billion donation from tech enthusiasts such as Elon Musk. Soon after, OpenAI’s ChatGPT hit the scene. This groundbreaking tool can generate sentences, text summaries, and even programming code, marking the beginning of a new AI era: the one of generative AI.

In 2022, generative AI models such as ChatGPT first caught the attention of the general public and truly broke into the mainstream. OpenAI presented DALL-E 2, an impressive image synthesis model that can generate images from text inputs. In August, text-to-image technologies became a hot topic, with Stability AI and CompVis launching Stable Diffusion 1.4, a powerful open-source image-synthesis model. In that same month, an AI-generated image titled “Théâtre d’Opéra Spatial,” created by Jason Allen, won a top prize at the Colorado State Fair fine arts competition. In late September, DALL-E 2 became publicly accessible, prompting a massive waiting list of excited users. In November, ChatGPT followed, amassing more than a million users in just 5 days. To put this in perspective, it took Netflix 41 months, Facebook 10 months, and Instagram 2.5 months to reach the same number of users.

With its ability to generate human-like text, generative AI such as ChatGPT is currently helping build new search engines, create personalized therapy bots, explain complex algorithms, and even write college essays. Text-to-image programs such as Midjourney, DALL-E, and Stable Diffusion are changing the game for art, animation, gaming, and architecture. With this new type of AI, the creative process is set to get a whole lot faster, as artists and designers have a new collaborator to augment their existing tasks and speed up the ideation and creation phase. In fact, most of the images in this book were created by generative AI, so make sure to look closely.

But it doesn’t stop there! Generative AI even has transformative capabilities in complex scientific disciplines such as computer science. Microsoft-owned GitHub Copilot, based on OpenAI’s Codex model, supports developers with code suggestions and automates up to 40 percent of their work.

With all these innovative use cases we can see already popping up, optimists believe that generative AI is the foundation for the future of creativity and complex sciences alike. So the evolution of AI surely doesn’t stop in the here and now. There are definitely some exciting times ahead of us!

The Evolution of Artificial Intelligence: Milestones on our Way to the Intelligent Machine

Buzzword Bingo Explained: From Machine Learning to Generative AI

Obviously, the basis of intelligence is—you guessed it—learning. Learning means we improve our performance in the future after we have observed and taken in some information about the world around us. This is no different for today’s AI. In the world of AI, learning often means that we have a vast collection of input-output pairs, from which an underlying function is derived so the model can predict the output for new incoming input. Sound complicated?

Let’s illustrate that with an easy example. Say we have a supercool collection of cute cat and dog pictures that show all different kinds of cats or dogs in them in all kinds of different situations. This is our input. As an output, we tell the model that these are either cats or dogs in the pictures. Now from this collection, we want to model to learn to recognize and distinguish cats and dogs. But not just those in our collection! This is no game of memory. That would be too easy. No; we want to feed the model a completely new cat image, and from this new input we want it to be able to predict that this is indeed a cat—and not our neighbor’s tiny chihuahua.

You may wonder, why go through all this hassle? Collecting all these pictures and tagging all the cats and dogs in there so the program can learn. Why not directly tell the machine all the steps to take to come to solve a problem or come to a conclusion? Just like math lessons, couldn’t we just simply let the program know all our math rules and functions so there is no more need for learning? Right . . . and wrong. This may work for simple tasks and problems. A lot of early AI research tried to teach the machine this way. And while they did have some success on smaller problems, they all failed miserably when it came to applying the models in the real world. And our world is messy, complex, and full of unknowns. So to be able to use AI for bigger and more complex problems, we need it to learn. Think about it. We simply cannot anticipate all situations the program will be in.

Consider a self-driving car. Now think about all cars on the road worldwide; the weather and road conditions; the cars’ wear and tear; and other cars, people, or obstacles on the road. Quickly we must realize that it is simply impossible to prepare an AI system for all possible situations. So it needs to learn. Also, there may be changes with time. Consider a program designed to predict the stock market. Now a global pandemic comes along. It needs to be able to adapt when conditions change from boom to gloom to make good predictions. So it needs to learn. And most of all, sometimes we have no clue ourselves! Just consider our cat pictures. While it may be easy for us to say that these are all cats, even the best programmers might find it hard to boil that down into an algorithm—unless it is self-learning. So in short, for AI to become truly intelligent, it needs to be able to learn!

This brings us to the next point. How does AI learn? You may have come across many of the terms here: machine learning, supervised learning, unsupervised learning, reinforcement learning, deep learning. Basically these are all different ways of a machine, learning.

Machine learning is the broadest concept of them all. It basically refers to the idea that our algorithm is able to learn. Remember how we said that for very simple problems, we can tell a machine directly what to do. So it would look something like this: “if this” . . . “then do this.” Clearly there is no learning from the computer’s side. It is us telling it exactly what we know. Now consider a more complex task, such as teaching our machine to tell the difference between cats and dogs. As you can imagine, it would be extremely hard and complex to narrow this down to an if-then sequence. So we want the machine to learn. Or put differently, we use machine learning. Although it may sound complicated, the term basically refers to a set of approaches that all share the idea that a machine will learn from the data it is provided to improve its performance on a given task. That’s it! That wasn’t hard after all, right? And you have all seen them before, too! Familiar examples are the Netflix recommendation system, Snapchat filters, Google Maps, and Spotify-generated playlists. All use AI models based on machine learning!

Supervised, unsupervised, and reinforcement learning are all forms of machine learning. The only thing that differs among these types of learning is the way feedback is provided to the system.

Very intuitively, supervised learning means we give our algorithm all the feedback we can give. In short, learning is supervised. What does this look like? Well, we simply use so-called “labeled” data. This labeled data contains both input and output, meaning the output is already known. Think about our cat and dog images. Instead of just feeding cat and dog images into the program, we also tell it if there is a cat or a dog in each picture by labeling the images as either “cat” or “dog.” So supervised learning is like having a teacher who shows you examples and tells you the correct answers. The more examples you see, the better you get at solving problems on your own. Once you get the logic, you receive new examples and solve them based on what your teacher taught you.

Clearly this labeling is time- and labor-intensive. So researchers tried to find ways around having experts painstakingly annotate data for supervised learning—and they have been successful! In 2022, a group of researchers were able to train an AI model called CheXzero to spot diseases on chest X-rays with medical reports that experts had written in natural language. And while you may think this hurts performance, it did not! In fact, the self-supervised model outperformed the supervised models with fully labeled data. This paves the way for self-supervised AI models that no longer need any data with explicit annotations and makes machine learning even faster!

Let’s move on to unsupervised learning, which is, as the name implies, learning without explicit feedback. No supervision. This means that the program is not given any labeled data. Instead, themachine is left on its own to group the unsorted information by finding similarities, differences, and patterns in the data. It’s likea detective trying to solve a mystery. Itdoesn’t have all the clues up front, so ithas to gather evidence and try to piece together what’s going on.

Consider our cat and dog images again. This time we do not tell which image depicts a dog and which a cat. The idea of unsupervised learning is that by carefully examining each image, the machine can identify the clues that separate cats and dogs, such as the length of their tails, the presence of retractable claws, or the number of whiskers. After analyzing the evidence, the machine can group the images into categories, but it can’t quite tell us if they are cats or dogs. Obviously this approach does not only apply to images of cats and dogs. In fact, this approach is often used in marketing, where marketers need to identify different customer segments. Through unsupervised learning, the AI can cluster customers together that share important features and are sufficiently dissimilar to other groups of customers.

Then there is reinforcement learning. You may have heard of this before. And you are right, reinforcement is a very common way we humans learn, too. When you think about reinforcement, think about rewards and punishments. We often use those in our day-to-day lives to reinforce positive behaviors in others while weakening negative ones. So we praise kids for eating their veggies while withholding the dessert if they don’t. How does this relate to AI? Well, we can provide feedback to our machine in the exact same manner so it may learn the best behavior or actions. For example, we can reward it with two points if it wins a chess game and deduct two if it loses. Similarly, we can add points if it correctly identifies our cats and dogs in the images.

The goal for the program then is to maximize the rewards and minimize the punishments. The machine achieves this by deciding which action prior to the reinforcement (positive and negative) was most responsible for it. It will then show those actions more often that led to a reward and reduce those that led to a punishment. So if it was a kid at the dinner table, it would eat those veggies to gather more praise and eat even more to make sure it no longer misses out on that yummy dessert!