Stephen Colbert doesn’t buy that machines can make art that feels human.
“The Late Show” host said on an episode of the “Possible” podcast published Wednesday that AI can mimic art but will always struggle to escape the “uncanny valley.” He asked whether AI-generated art will ever stop feeling “alien.”
True art comes from people because it fuses ideas with emotional experience, he said.
“Art is by humans for humans about being human. It’s not about ideas because ideas are constructs, and humans are not a construct,” Colbert said.
Humans “experience ideas and they have emotional responses to them,” he added.
Beyond art, Colbert also likened AI-mediated experiences to eating processed food: technically complete, but missing something vital.
“There is a growing suspicion that there are micronutrients of which we are not aware that we are robbed of,” he said.
Those “micronutrients,” he added, may lie in the very failures and imperfections of human connection. The flawed nature of human interactions is inextricable from our experience of the world, he said.
Colbert isn’t dismissing AI altogether. He said the technology could expand access to healthcare in underserved regions, where “a medical station manned by AI” might be better than nothing. But he remains skeptical of claims that machines can replicate human intimacy, such as in therapy.
Colbert did not respond to a request for comment from Business Insider.
AI and art
AI has already begun reshaping the art world.
Related stories
Business Insider tells the innovative stories you want to know
Business Insider tells the innovative stories you want to know
In Hollywood, studio giants have spoken publicly about their use of AI, pitching it as a way to both save money and make better films.
Lionsgate last September said it would work with Runway to experiment with its AI tools. AMC Networks followed suit this year. Amazon has talked about its use of AI to provide TV recaps and invested in Fable, a startup that lets people make their own TV shows using AI.
Netflix co-CEO Ted Sarandos said in July that the series “The Eternaut” was the streamer’s first to have shots fully generated by AI in its final footage.
“That VFX sequence was completed 10 times faster than it could have been completed with visual, traditional VFX tools and workflows,” Sarandos said about visual effects.
The release of OpenAI’s new image generator sparked a tsunami of Studio Ghibli-style memes and raised copyright concerns earlier this year. But OpenAI’s CEO, Sam Altman, said in April that AI art is a “net win” for society.
Altman said in an interview published on YouTube by the founder and content creator Varun Mayya that AI can expand creativity by lowering the barriers to entry.
Still, many worry about unauthorized or ethically questionable uses of AI.
In June, Disney and NBCUniversal sued the AI company Midjourney, saying it stole their famous characters, such as “The Simpsons” and “Minions.” Midjourney denied the claims in its legal response.
AI is central to a lawsuit that says Amazon MGM Studios violated copyright in its “Road House” remake and used AI to replicate actors’ voices to finish the film, in violation of labor contracts. Amazon said the suit was without merit and that the film didn’t use AI-cloned voices.
