Here is the audio, transcript, and summary of James Skinner’s talk titled “ChatGPT, AI, and the Crazy Future That Already Happened” at TEDxBorrowdale conference.
Listen to the audio version here:
TRANSCRIPT:
James Skinner – AI Educator
Ladies and gentlemen, welcome to today’s TEDx Talk, where we will embark on a thrilling journey to explore the astonishing world of artificial intelligence, or AI. My name is James Skinner, but the image you are seeing, the voice you are hearing, the text of my speech, the background music for this presentation were all created by AI and brought to life by cutting edge AI technology.
During our brief time together, we’ll dive into the fascinating realm of AI chat software and the crazy future that has just unfolded right before our eyes. Get ready to witness how AI has transformed the way we communicate, learn, and work, blurring the lines between human and machine intelligence. And fasten your seatbelts as we delve into the mind-boggling innovations and possibilities that AI has unlocked, forever changing the landscape of our lives.
Let’s begin.
Imagine walking into a bookstore and discovering that the number one bestselling book was written by a machine and not by a human. This is exactly what happened to me as I was preparing this talk during my recent stay in Tokyo. The seemingly impossible reality is our present, and it marks the beginning of a new chapter in human history as well as human-machine collaboration.
I’m sure you have many questions. You may be asking yourself, what does it mean to live in a world of intelligent machines? Can machines actually replicate or even surpass human creativity and originality? How will the widespread adoption of AI transform our economy, workforce, and job market? And most importantly, will my life be better or will it be worse? And what will the lives of my children be like?
Tsukuba International Expo 1985
In 1985, I was employed by the U.S. State Department and tasked with introducing and demonstrating U.S.-based AI technology to Japan and the world from within the futuristic U.S. pavilion at the Tsukuba International Expo. Our pavilion, adorned with cutting-edge technology, attracted visitors from all walks of life, including one very special guest in particular, Emperor Hirohito of Japan.
As the emperor’s arrival approached, the atmosphere was electric with tight security measures in place to ensure his safety. Secret service personnel with submachine guns were stationed strategically. Regular police officers monitored every major intersection within a five-kilometer radius, and a dense line of law enforcement created a human barrier at a 100-meter perimeter surrounding the pavilion.
Decked out in our bright red space-age jumpsuits, we awaited the emperor’s arrival with a mixture of anticipation and excitement. As he entered, it was my honor to demonstrate for him what at the time was a groundbreaking AI system capable of composing original classical music. The emperor, an accomplished scientist himself, watched with keen interest as the system set to work, composing a Baroque fugue right there in front of us.
When it had finished, it flashed the score on the computer screen and filled the pavilion with the melodic harmony of the AI-generated composition, which seemed to transport us to another time and place, leaving a deep impression on everyone present. As the final notes of the fugue echoed throughout the pavilion, Emperor Hirohito stared at the screen, utterly astonished.
His eyes widened, and in a moment of profound realization, he uttered a single word that encapsulated the marvel of the experience, haaaa. This encounter, a meeting between the emperor, an AI composer, and myself, illustrated and underscored the vast potential of AI, not just as a tool for crunching large quantities of data, but as something that could be creative in many of the same ways we are.
In short, machines could be intelligent. They could generate new things on their own. They could go beyond just doing what they were explicitly told to do.
[AI Machines Speaking]
Early AI focused on rule-based expert systems. The idea was that whenever you make a decision, you use a series of rules. For example, a doctor trying to diagnose an illness might ask if you have a fever. Then based on your yes or no answer, she might ask if you feel pain in your joints or if you have a rash. And following this decision tree of yes or no answers, the doctor will eventually conclude that you have the flu and should go home, rest, and get plenty of fluids.
Once we understood the decision tree in detail, we could program those instructions into a computer and young or inexperienced doctors could use the system to get a diagnosis as accurate as the best doctors in the field.
This approach, however, ultimately failed. The reason is simple. For the most part, we don’t know why we make the decisions we make. Why did you get the chocolate ice cream? It looked yummy. This is not something we can then program into a computer.
But from the late 1990s, as computer power advanced, we took a different approach. We decided not to try programming the computer at all. Instead, we would set up the computer so that it could learn. This ushered in the age of machine learning. The idea was to show the computer lots of data and have it make guesses. For example, we could show the computer a photo and ask it to identify the animal in the picture.
The computer might guess that it’s a dog. A human operator would then give it the correct answer, say in this case, a cat. The computer would then alter its algorithms on its own to try and be more accurate. This approach led to many miraculous things, such as recommendations on e-commerce websites, autonomous driving, identification of people you might know on social media, facial recognition, and more.
But now we have moved beyond that. We are now in an age of generative AI. We are now in an age in which AI can talk, and read, and write.
Pages: First |1 | ... | Next → | Last | View Full Transcript