Original article was published by Entre on Artificial Intelligence on Medium
The impact of AI-generated videos by AI startup Synthesia
Video for the most part has experienced very little to no disruption, the quality of videos may have increased but what does it look like when we completely rethink what videos should be like in the context of technology. This is an article that dives deep into the AI ecosystem: how it developed, how its disrupting video, and the startup leading the innovation Entre’s featured startup — Synthesia. Synthesia is disrupting video generation.
A Brief History of AI
First to understand just how truly revolutionary Synthesia is, we must understand the history of AI. The birth of AI was mainly in the 1950s, as with any new technology it sparked a lot of interest among those in the industry, and more broadly it sparked the imagination of the public as people began to make wild predictions about how the future would be, as with any revolutionary technology the fan fair soon waned along with it a lot of who once were optimistic funders of AI begin to pull back none more significant than the Government. Up until the early 21st-century innovation in the AI, space was little if any, the AI space would experience some peak interests and then go down again follow somewhat the same cycle for a while without any real progress. It was not until IBM in 1997 when they launched a research project that the original dream of AI was revived. The project spawned AI’s that were the first to beat Chess Grandmasters and as well as winning the quiz show “Jeopardy”. Moore’s Law meant that AI development could take place at exponential speeds. But for the most part, AI was a very niche industry although many predicted that eventually, it would make a significant change to the world it was thought that that was still a long way away as most of the applications of AI wherein Academia and building cool but globally unscalable products. In 2012 possibly the greatest development in AI took place. Researchers at Stanford and Google including Jeff Dean and Andrew Ng published their paper — Building High-Level Features Using Large Scale Unsupervised Learning, building on previous research into multilayer neural nets known as deep neural networks It would accelerate the pace of AI development and open up a new world of possibilities when it came to building machines to do work which until then could only be done by humans.
The most exciting company in AI and how disruptive they are
“Create your own AI video. As easy as writing an email.” — Synthesia CREATE is a powerful tool to create engaging video content without the need for actors, film crews, and studios. Synthesia is disrupting the very idea of what a human does in a video, their AI-generated videos are incredibly realistic and are a game-changer. They have the potential to one day be soo sophisticated enough to have entire movies be shot through AI actors. Synthesia may have passed the Turing test — Although possible it is very hard to distinguish between the AI and a real Human and this is revolutionary. It reshapes the way training material is produced, customer interactions with product info, and a variety of wide applications that have the ability to change the very way we interact with what are repetitive but very critical areas of business pertaining to human to human via video interaction.
How they do it
A recent breakthrough in the AI field that has allowed for what was a mere dream just a decade prior is the advent of the generative adversarial network(GAN). It is a class of machine learning frameworks designed by Ian Goodfellow and his colleagues in 2014. Two neural networks contesting with each other in a game (in the form of a zero-sum game, where one agent’s gain is another agent’s loss). Given a training set, this technique learns to generate new data with the same statistics as the training set. For example, a GAN trained on photographs can generate new photographs that look at least superficially authentic to human observers, having many realistic characteristics.
The impact of Synthesia
Imagine every dream we have ever had, imagine all the faces you have ever seen in your dreams. Most of them you may recognize, the same vaguely remember seeing, and then there are those faces we do not recall ever seeing. It would seem as if our brains created an individual within in our dreams as we sleep except that, that isn’t at all the case — humans are incapable of thinking of anew face, our brains cannot from the ground up synthesize a new face meaning that every face you have ever seen and thought of, is in part or holistically a face you have seen before but Synthesia has managed to create an AI that can do what humans cannot do. At the core of this sort of disruption is not a reason to fear that in the future we will not be able to tell apart reality and imagination as they intertwine but rather it paints a greater picture one which AI’s ability to exhibit just the right facial expressions and facial markers to embue another human with confidence o do whatever the AI was created to instruct, we as humans lack the capacity to replicate that same experience. As with any disruptive technology first, it is feared than it is adopted then it is revered like magic, then it is criminally under-appreciated. Is the world not better off with an AI that is able to cut costs without a compromise on quality and above all an AI that still remains motivated. Everything is scary until its not and I look forward to having my morning news be read by an AI-generated News presenter.