What is generative AI and what is its potential?
Deep Learning models have been around for a long time with researchers pushing boundaries every day and making breakthroughs almost every other week now.
One of those breakthroughs is generative AI. With OpenAI's Dall-E 2 and Stability AI's Stable Diffusion 2.0, any imagination can be visualised. All it takes is a sentence describing your imaginary photo.
This impressed a large audience quickly and easily. The news spread like wildfire and AI generated images were everywhere, from magazine covers to winning art contests:
Keeping aside all the controversial and ethical aspect to this piece of technology, let's talk about what it really means to our future. Where do you think this will lead us in the next 5 years?
Generative AI shouldn't be used only for imagery. Let's go back in time when technology was not advanced enough to record audio in its best form. Whenever I listen to some of my favourite old school melodies, I always think of how we can use generative AI to increase the quality of the audio by removing unwanted noise and generating audio filtered from the noise.
This will be an amazing application for this technology. In fact, this can be a great idea for a startup as well! Take any audio and enhance it to make it sound like never before. Talking about enhanced audio, there happens to be a speaker created in collaboration with Kanye West that uses AI to separate the vocals, bass, drums and instruments for you to remix the song in realtime on the speaker itself. Although not exactly the same application, to separate noise and tones you will need a similar tech that this speaker called STEM uses.
What's more? Imagine an actor who was injured during a film shoot. They would be unable to continue acting for the film. A possible workaround for this is to use a stunt double and morph the actor's face onto the stunt double's to make it look extremely realistic (For the sake of this article's topic, I'm not going to discuss the ethical implications of this technology).
Now, there seems to be so much more value to generative AI than just using it to create surreal images that do not exist. I gave examples from two completely different fields that one piece of technology can achieve, and this is the very beginning of generality. However, we have a long way to go for Artificial General Intelligence (AGI) and I will save that discussion for another day.
(The cover for this post is created by an AI)!
Thanks for reading!