When evaluating the most important innovations of a calendar year, it is often difficult to decide among a handful of equally worthy competitors. Not this year. Over the past 12 months, one category of technology has so often made the headlines and had such a significant impact on society that there is no doubt that 2022 will be remembered as the year that Generative AI stunned the world. I don’t just mean stunned the general public. Even longtime technologists and AI researchers like myself have been genuinely surprised by the speed and impact of recent advances.
Generative AI
So what is generative AI? It is a branch of artificial intelligence that allows computers to quickly and convincingly create original content ranging from images and illustrations to poetry, music, text, video, dialogue and even computer code. . The result is so impressive that it’s easy to imagine that we suddenly created sensitive machinery with a creative spiritbut it’s absolutely “not the case.
These systems are master imitators of human creativity. They have been trained on millions and millions of human artifacts such as documents, articles, drawings, paintings, films or anything that can be stored in large scale databases. These systems have no conceptual understanding of the information they process – to a computer they are just data models – and yet these generative AI tools can create original and impressive new pieces of content.
The underlying technology has been around for a few years, but it wasn’t until 2022 that the systems reached a level of maturity where they could be made public for widespread use. For example, this year the world was flooded with AI-based image generation tools, including very popular systems such as DALL-E 2, Stable Diffusion and Midjourney. They are so easy to use and can produce amazing results so quickly that they sent shockwaves through the art community. The entire industry has been disrupted by these new AI competitors that can offer a wide range of artistic options to choose from in minutes based on a simple text prompt – often for free.
It’s no wonder people are flocking to generative AI tools. ChatGPT, which we discuss in more detail below, has reached one million users in one week. By Comparison, it took ten months for Facebook and 2.5 months for Instagram to reach the same milestone. According to Open AIover 1.5 million users use DALL-E to create over two million images per day. Steady broadcast has more than ten million daily users, and Mid Road has over two million members.
To give you an idea of just how fast and flexible these generative illustration tools are, I hopped into Midjourney and asked him to create a postage stamp to commemorate the awesome power of generative artwork. AI in 2022. To do this, all I did was give the system asked, “Imagine a postage stamp with a robot holding a paintbrush”, and within 60 seconds the system gave me this original image.
Considering I put in about two minutes of effort and got the awesome result above, it’s easy to see why commercial artists are concerned about their livelihood. And the systems keep improving, allowing users to give feedback and request changes to every item generated. So, I spent another few minutes and asked for the image to be more colorful, for the robot to have a smile, and for me to add “2022” to the image. I thought this might make for a more festive postage stamp to commemorate this remarkable year. In 60 seconds, that’s what Midjourney produced.
To be honest, I prefer the first stamp. I say this because the robot has a slightly guilty look on its face, like it knows it’s an impressive performer, but also that we humans should be careful in our celebration of this milestone.
Not just art, but also text
After all, generative AI systems will disrupt the human workforce that goes far beyond the visual arts. Indeed, in 2022, generative AI systems have also become incredibly good at composing texts on almost any topic. Known as Large Language Models (LLM)these systems have reached a level of maturity such that they have now been deployed for beta use by millions of people around the world.
The most impressive text generator to be released this year is ChatGPT of Open AI. You can enter simple text prompts or engage in conversational dialogue, and it instantly produces consistent messages and most of them accurate results. This created a wide range of reactions ranging from excitement to genuine concern.
The excitement is that large language models will change the way we interact with computers at all levels, ushering in a new era in which we can communicate through fluid, natural dialogue. At the same time, many people fear that systems like ChatGPT are so good at writing consistent documents and computer code that they will quickly rival human professionals for many tasks. There is also the concern of educators that tools like ChatGPT will make it impossible to assign take-out essays. After all, with ChatGPT, a student can simply have the computer write it down for them – and because it’s original content, it may be impossible to determine that cheating has occurred.
Rather than tell you why people are worried that generative AI systems will kill human jobs, let me ask ChatGPT to tell you. And to prove just how remarkably powerful and flexible these systems are, allow me to enter this simple creative prompt: “write me a Dr. Seuss-style story about an AI system that writes creative items for people, taking our jobs away from them.” Here is what the system produced in seconds:
Once upon a time in a land not so far away
He lived smart and intelligent AI
This AI system was named Wizard
And he could write articles with great perseverance
He wrote about science and technology
And even dabbled in poetry
People marveled at his clever words
And the way it could make their thoughts fly
But over time people started to get scared
This assistant would take their jobs away
They feared that he wrote better than them
And leave them unemployed, doing nothing all day
Subscribe to get counterintuitive, surprising and impactful stories delivered to your inbox every Thursday
But the assistant just wanted to help
It didn’t mean to cause strife
She wanted to share her creative talents
And bring joy to people’s lives
So people learned to kiss
The help that the assistant could provide
They used it to improve their own writing
And together they reached new heights
The end.
Sure, the story above isn’t as clever or charming as something Dr. Suess would have written, but the fact that a computer instantly created a story on demand that is so coherent is frankly breathtaking. And we are still in the early days of generative AI. Next year around this time, I’m sure the ChatGPT system will do a much better job. Moreover, it is not only text and images that these technologies learn to generate. New systems are being developed around the world that can create music and videos with impressive results. As we look to the future, generative AI will be a critical element of the metaversecreating and deploying fully immersive experiences that look and feel authentic.
A false impression of precision
Personally, my biggest concern with generative AI systems is that we humans can assume that their output of information is accurate because it comes from a computer. After all, most of us grew up watching shows and movies like star trek where characters verbally ask computers for information and instantly get accurate and reliable results. I can even hear Captain Picard in my head barking a command like, “Computer, estimate how long it will take us to catch up to this space probe.” And an authoritative answer comes back. Everyone believes in it. After all, it’s from a computer.
But here’s the thing: Generative AI systems are trained on massive sets of human documents that aren’t fully verified for accuracy or authenticity. This means that training data may include material filled with misinformation, misinformation, political bias, or social bias. For this reason, ChatGPT and other systems include disclaimers such as “May sometimes produce incorrect information” and “May sometimes produce harmful instructions or biased content”. It’s good that they tell you this upfront, but I’m afraid people will forget the disclaimers or don’t take these warnings seriously. These current systems are not factual databases; they’re designed to mimic human responses, which could easily mean mimicking human flaws and errors.
Whether you think this is a good step for humanity or a deeply concerning step, you have to agree that generative AI technology is poised to change society in significant ways. And while 2022 has been a year of stunning progress in this area, we have no reason to believe it will slow down anytime soon. For example, Open AI is already working on the next version of ChatGPT using a more advanced language model called GPT4, which should be available next year. Who knows, maybe next December this whole year-end article will be AI-generated.