The year 2023 marked the explosion of generative AI across text, images, video and many other applications. By March of ‘23, the creative and marketing fraternity got extremely excited by the amazing possibilities that generative AI brought to bear. However, I believe it was the tip of the iceberg with many people playing around with the tools available, using it for ideation and writing copies for social media posts and the like. But by the time we reached the end of 2023, most of these models had evolved significantly and are now a completely new generation of what we saw six to nine months back, in terms of capabilities.
As with any technology, there are good and bad sides, and good and bad applications. AI brought with it a moral dilemma with lots of tech leaders asking for pause on the evolution of these learning models. There have been cases pertaining to IP protection but a lot of that seemed to be getting settled as the year progressed. The rise of deep fakes and other grey area usages of AI have also emerged, which is a serious concern as far as reputation management, security and even national interest is concerned. With the advent of 2024, things are going to get really crazy.
Decoding ChatGPT, Midjourney & More
While everybody is waiting for newer models of the likes of ChatGPT, Bard, Dall-E or Midjourney, there are many other aspects of AI that would most likely explode in the year 2024.
Some of the areas that I am personally extremely excited about are with respect to AI’s ability to understand complicated user prompts involving a number of subjects, images, colours, textures, scenery, as well as emotions, to be able to generate the exact image that marketing professionals seek. This has huge implications on productivity, ability to quickly create campaigns, as well as the associated savings in licensing costs or custom photography.
The second area is the ability to create end-to-end video assets using not just textual commands and input images, but also giving very specific directions like the areas to animate, motion paths, guides, and the ability to even do spot corrections in motion videos.
When Artificial Intelligence Gets Realistic
The area of audio leading up to custom voices and custom avatars, making them completely believable and realistic with hand gestures, body gestures and emotions, would also likely see a lot of development. This again has implications on the learning industry, besides in corporate presentations and videos.
Overall, I believe that all this will create a sequential workflow, where people will be using not just one or two but multiple AI tools for very specific purposes. For example, upscalers are no longer about simply increasing resolution, but about adding significant details to the image. Magnific AI is an amazing tool that has been recently launched by a small startup but it is seeing a tremendous amount of traction. In product photography, the ability to input 2D images to get 3D objects would see great development in the year. This would allow people to do table-top shots and create product packaging completely using AI without the need for expensive studio shoots.
What's in for 2024?
The holy grail will be when full 3D scenes can be created for commercial use. The ability to scan a room or a scenery and be able to create 3D motion paths through and around these objects would give creative professionals ultimate freedom. All in all, 2024 is likely to see accelerated development and even more amazing use cases coming out of the AI community.