The CEO of one of the world's leading artificial intelligence (AI) image and video enhancement software companies pushed back on concerns surrounding AI's impact on art and entertainment employment, claiming recent advancements would benefit creators by removing technical challenges.
"There's a common fear that AI will replace humans, but we believe that we'll instead see an explosion of AI-assisted human creativity and productivity in the next few years. People create art to tell stories to other people, and AI can never replace that," Topaz Labs CEO Eric Yang told Fox News Digital.
A March economic research report from Goldman Sachs suggested that AI has the potential to automate up to 26% of employment tasks in media, design, arts, entertainment and sports.
Meanwhile, the discourse from artists on generative AI has been mixed. Some have called AI art and photo generators "anti-artist," noting that the models rely on already created human artwork for training data obtained without consent or compensation.
Others have expressed concern that AI's ability to create work in the vein of other artists swiftly could make the general public perceive that photography and painting are easy to assemble and replicable. Often, artists' fears are levied at generative AI models, like MidJourney and Dall-E, as opposed to image or art enhancement tools.
But Yang likened new AI enhancement and generation software to prior concerns around tools like Photoshop. In 1990, when Adobe announced the photo-editing software, many artists expressed concern it would lower the value of photographs and put artists out of a job.
"Photoshop didn't replace photographers, it just gave them superpowers," Yang said. "Similarly, generative AI will enhance the talent of human artists rather than replacing it. With AI, creators can fully concentrate on their vision instead of being bogged down by technical details. We will soon see some of the most imaginative work ever produced as a result of combining AI tools with human taste and storytelling."
The Topaz Labs CEO also discussed how deep learning models for image enhancement were first proven commercially viable with great results and acceptable speeds in 2018. Since then, AI tools have steadily replaced previous tools for challenging tasks, such as noise reduction, image sharpening and upscaling.
In 2023, those forms of image "enhancements," wherein an editor improves what was already there, have extended from enhancement to "generation," the process of creating something out of nothing. Yang predicted that the AI generation would experience a similar trajectory to AI enhancement over the next five years.
"Traditional editors force you to lay out the exact set of tools and instructions to achieve a desired outcome. AI software is different because they're ‘descriptive:’ you specify the desired end result and the software figures out how to get there," Yang said. "It's like moving from being the painter to being the art director: you're still in control but you don't have to brush every stroke yourself. This will affect every feature in current editors, but also include new features that we haven't even thought of yet."
Yang noted that these new software options would increase accessibility and efficiency in video and photo editing. However, he surmised that the "real value" of this descriptive AI editing system will lead to an unshackling of human creativity from the technical limitations of prior "crude tools."
"Let the software worry about specific technical details, and you can focus entirely on achieving your artistic vision. We started creating better photos after digital replaced film, and we will create better art after AI tools replace Photoshop," Yang added.
Looking forward, Topaz Labs Founder and Research Advisor Albert Yang said the trend of AI replacing hand-crafted and time-consuming algorithms for image and video enhancements would continue much faster over the next ten years. He also suggested that AI will continue to open the door to previously impossible methods for video and image manipulation, an aspect highlighted in recent years by evolving text-to-image generation and deepfake technology.
Deepfake technology has improved dramatically over the last several years. Experts have warned that deepfake tech could cause a wide range of problems, including increased scams, disinformation campaigns and fake news stories.
A recent Yale survey asked students whether art was generated by a human or by AI. Those surveyed were asked to judge nine art pieces, which included a mixture of AI and human-generated pieces. On average, the undergraduate students surveyed answered correctly 54% of the time, translating to five correct guesses out of nine artworks.