What is generative AI?

- It might seem incredible, but we're now in an age where in a matter of minutes I can create a well-written blog post to promote my business, design some unique artwork options to accompany it, generate a nice background music track from my podcast, and put together some snappy social media posts to promote it all. All of this is thanks to the power of generative artificial intelligence. Generative AI is a type of artificial intelligence where people can prompt a computer to create new, original content like text, images, video, audio, and more. This is all accomplished via a special type of machine learning. Let's talk about how this works. When an AI model is originally created, it's fed training data, lots and lots of training data. Images, text, voice recordings, it depends on the type of AI being trained. But in each case, the AI learns through the input of massive amounts of information. Once the AI model has been trained, people can ask it to make something new. In many cases, this happens by writing a short piece of text called a prompt. The AI model then attempts to create a novel artifact, an image, a collection of text, or an audio or video recording. If done well, the results reflect what the prompt asked for, whether that's a realistic photo, a futuristic image of a robot, a poem, or a set of instructions. Let's take an example of image-based generative AI. If an AI model is fed thousands of images of a horse and is taught by people that each of those images is a horse, it eventually learns what a horse should look like. Then when a person asks it to create a horse, it generates a completely authentic looking, unique image of a horse that doesn't actually exist on earth. This alone is pretty neat, but as you can see here this can get really intricate. Generate a blue horse crossing a river in the darkness under a full moon in the style of Van Gogh. Pretty amazing, right? And if that doesn't impress you, take a look at this website, which is full of people's faces who don't actually exist. Things work similarly for text-based generative AI. Let's take the example of ChatGPT, which is trained on a massive dataset of human generated text from the internet. When a person provides a prompt, ChatGPT uses patterns and relationships it learned from the training data to generate a unique response. If we go back to my examples from earlier, let me show you the type of prompt that generated my personal blog post. Notice how I define the tone as well as the main details I want to emphasize. And here's my prompt for my snappy social media post. Pretty straightforward, but it's a great way to mass create useful social content. There's a lot of excitement about this tech. Already creative pros use image-based generation tools to brainstorm and ideate. Professionals use text generation tools to streamline all sorts of writing. And people with communication challenges use tools to help them overcome their impairments. At the same time, some people have fear and skepticism about this tech. Some are scared that synthetic media like deepfakes could deceive entire populations, or that jobs traditionally done by people may disappear or change. There are also concerns around ethics, copyright, and bias. Positive, negative, and in between, all of these reactions are valid. This technology is evolving rapidly, and we're all on this journey together, but we each have a say in where this journey takes us. Because after all, for generative AI to be used effectively people must be involved every step of the way. You need people to train the AI and to provide valuable feedback to improve and maintain its accuracy. You need people to create well-crafted prompts. And you need people to thoughtfully check, analyze, and edit the results. Yes, generative AI will certainly change how many of us live and work, but the collective spark of our humanity is something we can't underestimate. We need clever, creative, collaborative people to make it all work. It's up to each of us to learn about the creative possibilities, the issues, and the ethics of this powerful technology. And that's what we'll spend the rest of this course exploring.