It's literally Her and ScarJo is not happy (and with good reason)
Sam Altman and OpenAI just proved Hollywood's fears of AI are warranted.
Last week, on Monday, May 13th, OpenAI streamed a live “Spring Update” where the company introduced “real-time conversational speech” (watch here; the voice demo starts at 09:43).
The demo was undeniably impressive. The ChatGPT chatbot sounds far more human and emotional than anything I’m used to from Siri, Alexa, or Google Assistant. The cadence of the female voice is much more human-like, changing tone and pitch, inserting um’s and ahh’s for effect, and sounding distinctly flirty at times, with more than passing similarity to Scarlett Johansson’s voice as the sentient AI “Samantha” in the Spike Jonze movie “Her.”
Turns out, it is maybe more than a passing resemblance. Yesterday, Monday, May 20th, Scarlett Johansson released a statement to NPR in which she detailed that she had been approached by Sam Altman to be the voice of ChatGPT 4o by Sam Altman, an offer that she (unsurprisingly IMHO) declined “for personal reasons.” (you can read her full statement here)
OpenAI had clearly tried to get ahead of Johansson by putting out a statement saying that “Sky’s” voice wasn’t modeled on Johansson’s voice and that we’re all listening to it wrong if we think it is. In an interview with The Verge following the event, OpenAI's CTO Mira Murati expressly denied that the voice was modeled on Johansson’s: “Someone asked me in the audience this exact same question, and then she said, ‘Ah, maybe the reason I didn’t recognize it from ChatGPT is because the voice has so much personality and tonality.’”
The fact that Sam Altman himself Tweeted (I will never call it X, not sorry) “her” back on May 13, the same day of the Spring Update event, seems to have slipped Murati’s and everyone else in OpenAI's minds and could now put him in serious legal trouble.
Clearly, this story is still breaking, and more will unfold over the coming days. But, I think it’s safe to say that, at the very least, this is a massive self-own for OpenAI in general and potential reputational damage for Sam Altman in particular.
Sam Altman has revealed himself to be pulling a trick from the Silicon Valley playbook—"We love creators and respect their talents and art as long as they do exactly what we want"—far more than many of us (well, me) thought or hoped. It now appears that when he was rebuffed by one of the biggest movie stars in the world (for completely understandable reasons), he seemingly copied her voice anyway and called it Sky instead of Samantha—which, really, you couldn’t at least pick a different first letter or add a few more?
Regardless of how this situation resolves itself, I think it’s safe to say that this will hardly assure anyone working in any creative industry that OpenAI has their best interests at heart. And to be blunt: This is such a huge fuck you to actors and Hollywood that I wouldn't be at all surprised if Apple, with all of its investment in streaming and entertainment, is potentially reconsidering its options for a partnership with OpenAI right now. After all, a few weeks ago, Apple was vilified for releasing an ad that literally crushed the timeless tools of the creative industry, so maybe they take a beat before allying with the company that is literally coming for our voice and likenesses, too.
I suspect that more than a few people are wondering whether the members of the OpenAI board who called for Altman’s firing actually had a few solid reasons for doing so.