OpenAI and Sam Altman promise to change the world, but should we trust people who keep making dumb mistakes, like they did with Scarlett Johansson?
Jason Merritt/Getty Images; Kevin Dietsch/Getty Image; BI
OpenAI's fight with Scarlett Johansson isn't just a PR disaster (and a big one at that). It also reveals a troubling attitude about people's intellectual property — and even their voice.
Most of all, it shows there's just really, really bad judgment going on at the highest levels of Sam Altman's company.
Let's say ScarJo had actually agreed to do the voice of "Sky" for ChatGPT instead of Altman's company hiring someone else to play the breathy, flirty female part. In this scenario, the voice would be not just vaguely reminiscent of the movie "Her" — a winky reference for those in the know — but literally an exact imitation of the movie, with Johansson playing the starring role.
But … stop for a second and imagine how truly stupid that idea would be from the start.
First of all, getting one of the highest-paid Hollywood actors to endorse your product is expensive. And getting her to reprise a role from 2013 is for what, exactly? Just a nerdy joke? Not even that many people saw "Her" (and I do worry our tech overlords forgot the ending).
Also, consider that maybe a voice assistant shouldn't necessarily be a sexy-breathy-flirty American female voice (yes, there are other voice options, but OpenAI chose "Sky" for the demo).
It's giving "let that sink in." It would be funny for a Super Bowl ad — not to develop into the flagship product that you're also claiming will change the world.
What actually happened, of course, was way worse. Scarlett Johansson declined when Sam Altman asked her to lend her voice to ChatGPT back in September when the SAG strike was going on, CNN reported.
Meanwhile, OpenAI hired a voice actress who the company said wasn't intended to be a voice mimic. But everyone immediately noticed that the "Sky" voice that ended up on ChatGPT reminded them of ScarJo.
And then Sam Altman, in one of the greatest self-own moves of the generative AI era, tweeted out "her" during the product demo last week. Doesn't that make it seem like OpenAI did want to evoke Johansson's character from the movie?
The company said in a statement that the voice wasn't intended to mimic Johansson's. Altman said in a statement to Reuters that the actor used for "Sky" was cast before he contacted Johansson to ask to use her voice. "Out of respect for Ms. Johansson, we have paused using Sky's voice in our products," Altman said. "We are sorry to Ms. Johansson that we didn't communicate better."
Now, Altman and OpenAI are tangling with an immensely popular celebrity, adored by a nerd army for her superhero role in the wildly popular Marvel franchise, with a husband whose job is to make fun of things in the news on TV each week, and a history of litigation against a big company.
Of course, it's also hitting a nerve because now there's a beautiful and popular face attached to something that the general public has been slightly nervous about: "Is AI going to steal our content and replace us?"
Altman tweeted himself into tech's biggest PR disaster since Facebook Beacon — completely avoidable. The OpenAI team used a Scarlett Johansson sound-alike voice — also completely avoidable.
It almost makes you wonder: Are these really the people we should believe are such geniuses and have such great judgment that they should be in charge of this potentially life-changing technology?
Just last week, several high-level OpenAI employees, including Ilya Sutskever, also a board member, left the company. This comes six months after what OpenAI and Microsoft employees internally referred to as the "Turkey-Shoot Clusterfuck" — the Thanksgiving week firing of Altman and his eventual return the next week. The details of what went down first were eventually revealed not to be a valiant fight of good and evil over controlling an all-powerful AGI, but apparent run-of-the-mill workplace gossip and alleged bullying.
This is all enough to make even the most optimistic ethical accelerationist pause and wonder, uh, what the hell is going on at this clown show? If AI is as powerful and world-altering as OpenAI is telling us it will be, are we not supposed to be mildly concerned that the people in charge keep stepping on rakes?
