OpenAI today announced GPT-4o, a new AI model that will be available to both free and paid users. Among its many improvements (faster response times, improved memory capabilities, better image analysis) is a conversational voice that does its best to sound like a real human being. He laughs, jokes, maybe flirts a little. “It looks like AI from the movies,” OpenAI CEO Sam Altman wrote in a blog on Monday. mail. “I’m still a little surprised that it’s real.”
To be honest, it looked like the AI from a particular movie: His, Spike Jonze’s 2013 sci-fi film that correctly foresaw a future in which AI relationships could easily replace human connection; Well, that’s what it felt and sounded like. In the demo, ChatGPT’s voice is remarkably similar to that of His star Scarlett Johansson. In case there was any doubt about the reference point, Altman tweeted “her,” just one word, shortly after the event.
His It’s a great movie. His vision of AI is surprisingly nuanced, and his depiction of the techno-human relationship at its core leans more toward utopia than instinctive skepticism. Still, a plea to anyone trying to manifest Jonze’s world, or that of any sci-fi touchstone, for that matter: look at it just one more time. All the way through. Just to make sure we’re all on the same page about the future we’re headed toward.
As my colleague Kate Knibbs recently pointed out, AI assistant Samantha at His It’s not malicious. There is no need to take the easy and beaten path of turning against humanity. It doesn’t even separate him from the rest of society; AI companions are so normalized in Jonze’s imagined future that no one bats an eyelid when Samantha’s user, Theodore, takes Jonze as his date on a double date.
It’s pretty easy to see why His It has a lot of appeal for AI companies. At a glance, it has all the benefits of conversational general AI and none of its drawbacks. (In particular, as Knibbs also mentions, none of the job displacements or economic disruptions that AGI portends.) But the fact that the inhabitants of the world of His Not having problems with the AI company does not mean that it is an unlimited good. Their relationships with AI are easy, safe, but also false. Samantha exists to meet Theodore’s needs; She is a dynamic that allows her to receive without giving, to have the constant security that she is understood without having to make an effort to understand another person.
It’s not until Samantha leaves—in His, the world’s AIs disappear to a higher plane of existence, an outcome that would surely alarm OpenAI investors: Theodore facing his own messy human relationships. He writes a letter to his ex-wife. He watches a sunrise with his neighbor. These are simple human acts, deferred due to enabling AI. Roll credits.
Honestly at least His offers a relatively sunny version of the future to hold on to, even if we disagree about what should be made of it. It’s one of the least offensive examples of the tech billionaire class’s sci-fi yearning. Elon Musk has alternatively described the Cybertruck as “designed for Bladerunner (sic)” and “what Bladerunner (sic) would have led to.” As Max Read has observed better than ever, this is wrong on a staggering number of levels, one of which is that the future of Bounty hunter It’s not something anyone should aspire to.