The quantum leap in generative Artificial Intelligence video

Artificial Intelligence is turning everything on its head. SORA, the new generative AI video tool, has been in the spotlight for weeks now—because every week there’s a new breakthrough. With natural language prompts, you can get incredibly realistic, long, high-quality videos. In this video you can see a compilation of demos to discover what it’s capable of doing. Keep reading on to find out where it still fails and how the combo of human creativity + AI can produce more than just interesting results…

You’re speechless, right? Even so, the first reaction we got was cooler. You see, at Equilateral we have Mrs. and Mr. Frame, who head a super-efficient quality and supervision department. That’s how we ensure that the first version of any audiovisual we deliver to the client is already close to perfection. Their well-trained eyes immediately saw that the cat had three forepaws…

…and that in a couple of frames, the corners of a book were bending in a completely unnatural way…

And then, in successive viewings they detected more “glitches in the Matrix”; in other words, what most professionals in the field detected: that the laws of physics in motion are still unattainable for artificial intelligence, although…only sometimes.

And then, when we were boasting about human superiority over machines, a chill ran through our bodies from our cerebellum to our toes: we realized that AI had taken a titanic leap in just 12 months. In just one “human” year, which seems to be the equivalent of 20 years for AI (in terms of what it has done), we have gone from coming across some disturbing, implausible and poor quality videos such as the ones generated by Stable Diffusion in 2023 (like the famous video of Will Smith eating spaghetti), to this hyperrealistic madness that, okay, has its flaws but is very powerful and holds a lot of potential.

This fast-paced evolution has significantly influenced our industry, and we have the best example in the upper echelons of Hollywood, where a producer has put an investment of 200 million dollars on hold for future productions after seeing the SORA demo. From his point of view, it is no longer necessary to arrange and transport a film crew when he can generate images without even leaving the studio. The rules of the game are definitely changing.

At Equilateral, we highly value the experience and creativity that our team can contribute to any audiovisual project, since human judgment and care are irreplaceable. So, in the end AI is a good ally for us when it comes to streamlining routine tasks, helping us generate images that don’t exist and giving extra quality to our productions.

We are early adopters and, over the last few months, we have been studying and implementing AI tools to create images that we haven’t been able to photograph or buy in stock libraries, or which simply don’t exist. It has also helped us gain new perspectives on our own images, or even to unblock brainstorming processes on beginning a script. The possibilities are endless, and every day we are discovering new features that are increasingly more awesome and precise.

Far from “taking our jobs away,” these tools are opening up new job prospects to us. There are a lot of companies that already have talented people who can act as a bridge between human language and artificial language. These prompt-engineering professionals know how to formulate the most precise and effective instructions to obtain optimal outcomes after having studied computer language through sheer practice within these systems or through ad hoc training.

Nevertheless, there is a much more imminent and very real danger with these kinds of artificial intelligence. We mustn’t forget that they are collaborative media that require our human intervention to provide information of interest and real data. The security of the documentation we provide may get compromised, not to mention our ethical principles. So, we should not let Artificial Intelligence compromise our fundamental rights as human beings, professionals and artists.

Having come this far, we are now wondering what the next steps will be in our audiovisual industry. Shall we attempt to put restrictions on its evolution? Will it replace any of our colleagues in the industry? We could ask it! But perhaps the machine’s artificial cunning will only give us an answer that’s deliberately confusing. I wrote that as a joke, but on re-reading it I see the sense in it; excuse me for being dystopian… But now if I take a more practical and optimistic attitude, I think the appropriate ethical limits will end up being put in place, and it will be a very powerful, useful tool.

What we must be clear about is that any company investing in AI as a back-up for its in-house creative department will be able to count on a professional duo capable of streamlining the creation processes and enriching the human-machine relationship.

We can see the perfect example in this video creation by the artist Don Allen Stevenson III (formerly in Dreamworks):

If you had time to read the description of the video on YouTube, you’ll have realized that the creator’s stance is quite positive regarding this new creative period. He himself defines it as “a new era of creative storytelling.”

By directing AI well and adding an immense pinch of creativity and human experience, fantastic results can be generated. A great many more audiovisual pieces are being envisioned, making creative dreams that have been difficult to crystallize until now come true (imagine the 3D animation costs that this short video by Don would have had!) and a horizon with surprises that we cannot even conceive of right now.

But we should never forget that it must always count on supervision from human judgment and our natural predisposition for creative diversity. This way is a must for communication experts because they fear that Artificial Intelligence will lead us to a kind of uniformity and homogeneity of audiovisual content, since at the moment, everything is being drawn from the same sources.

Stay tuned!  😉

We’ll continue to report…

Toni Hernández González. CTO and Post Production Director

Xavi Moreno Simón. Project and Content Director


More ideas...