More stories

  • in

    The Voices of A.I. Are Telling Us a Lot

    What does artificial intelligence sound like? Hollywood has been imagining it for decades. Now A.I. developers are cribbing from the movies, crafting voices for real machines based on dated cinematic fantasies of how machines should talk.Last month, OpenAI revealed upgrades to its artificially intelligent chatbot. ChatGPT, the company said, was learning how to hear, see and converse in a naturalistic voice — one that sounded much like the disembodied operating system voiced by Scarlett Johansson in the 2013 Spike Jonze movie “Her.”ChatGPT’s voice, called Sky, also had a husky timbre, a soothing affect and a sexy edge. She was agreeable and self-effacing; she sounded like she was game for anything. After Sky’s debut, Johansson expressed displeasure at the “eerily similar” sound, and said that she had previously declined OpenAI’s request that she voice the bot. The company protested that Sky was voiced by a “different professional actress,” but agreed to pause her voice in deference to Johansson. Bereft OpenAI users have started a petition to bring her back. More

  • in

    What We Lose When ChatGPT Sounds Like Scarlett Johansson

    OpenAI has good reason to aim for a bot voice à la the one in “Her.” But that film was about relationships. What does this real-world turn say about us?When Spike Jonze’s romance “Her” was released in 2013, it sounded both like a joke — a man falls in love with his computer — and a fantasy. The iPhone was about six years old. Siri, the mildly reliable virtual assistant for that phone, came along a few years later. You could converse in a limited way with Siri, whose default female-coded voice had the timbre and tone of a self-assured middle-aged hotel concierge. She did not laugh; she did not giggle; she did not tell spontaneous jokes, only Easter egg-style gags written into her code by cheeky engineers. Siri was not your friend. She certainly wasn’t your girlfriend.So Samantha, the A.I. assistant with whom the sad-sack divorcé Theodore Twombly (Joaquin Phoenix) fell in love in “Her,” felt like a futuristic revelation. Voiced by Scarlett Johansson, Samantha was similar to Siri, if Siri liked you and wanted you to like her back. She was programmed to mold herself around the individual user’s preferences, interests and ideas. She was witty, and sweet and quite literally tireless. In theory, everyone in “Her” was using their own version of Samantha, presumably with different names and voices. But the movie — which I love — was less the tale of a near-future society, and more the coming-of-age story of one man. Theodore found the strength to return to life in a brief, beautiful relationship with a woman who fit his needs perfectly and healed his wounds.It was thus a tad jarring to hear the voice of the virtual assistant in last week’s announcement of the newest version of ChatGPT, probably the best known artificial intelligence engine in the very real world of 2024. Among other things, the new iteration, dubbed ChatGPT-4o, can interact verbally with the user and respond to images shown to it through the device’s camera. Those who watched the live demo from OpenAI, the company that makes ChatGPT, were quick to note that she sounded a whole lot like Samantha — which is to say, like Johansson.Mira Murati, OpenAI’s chief technology officer, told The Verge that the resemblance was incidental, and that ChatGPT’s nascent speech capabilities have used this voice for a while. But once you hear it, you can’t unhear it.Those who watched the live demo from OpenAI, the company that makes ChatGPT, were quick to note that she sounded like Samantha.Warner Bros. PicturesFurthermore, OpenAI founder and chief executive Sam Altman has professed his love of “Her” in the past. Following the announcement, he posted the word “her” to his X account. And on his blog post about the news, he wrote, “It feels like A.I. from the movies; and it’s still a bit surprising to me that it’s real. Getting to human-level response times and expressiveness turns out to be a big change.”We are having trouble retrieving the article content.Please enable JavaScript in your browser settings.Thank you for your patience while we verify access. If you are in Reader mode please exit and log into your Times account, or subscribe for all of The Times.Thank you for your patience while we verify access.Already a subscriber? Log in.Want all of The Times? Subscribe. More

  • in

    Striking Writers Are Worried About A.I. Viewers Should Be, Too.

    A.I. screenwriting, a point of contention in the Writers Guild strike, may not yet be ready for prime time. But streaming algorithms and derivative programming have prepared the way for it.Television loves a good sentient-machine story, from “Battlestar Galactica” to “Westworld” to “Mrs. Davis.” With the Writers Guild of America strike, that premise has broken the fourth wall. The robots are here, and the humans are racing to defend against them, or to ally with them.Among the many issues in the strike is the union’s aim to “regulate use of material produced using artificial intelligence or similar technologies,” at a time when the ability of chatbots to auto-generate all manner of writing is growing exponentially.In essence, writers are asking the studios for guardrails against being replaced by A.I., having their work used to train A.I. or being hired to punch up A.I.-generated scripts at a fraction of their former pay rates.The big-ticket items in the strike involve, broadly, how the streaming model has disrupted the ways TV writers have made a living. But it’s the A.I. question that has captured imaginations, understandably so. Hollywood loves robot stories because they make us confront what distinguishes us as human. And when it comes to distinguishing features, the ability to conjure imaginary worlds is simply sexier than the opposable thumb.So the prospect of A.I. screenwriting has become potent, both as threat and rallying cry. Detractors of the striking writers taunted them on social media that software was going to horse-and-buggy their livelihoods. Striking WGA members workshopped A.I. jokes on their picket signs, like “ChatGPT doesn’t have childhood trauma.” (Well, it doesn’t have its own. It has Sylvia Plath’s, and that of any other former unhappy child whose writing survives in machine-readable form.)But it shouldn’t surprise anyone if the TV business wants to leave open the option of relying on machine-generated entertainment. In a way, it already does.Not in the way the WGA fears — not yet. Even the most by-the-numbers scripted drama you watch today was not written by a computer program. But it might have been recommended to you by one.Algorithms, the force behind your streaming-TV “For You” menu, are in the business of noticing what you like and matching you with acceptable-enough versions of it. To many, this is indeed acceptable enough: More than 80 percent of viewing on Netflix is driven by the recommendation engine.In order to make those matches, the algorithm needs a lot of content. Not necessarily brilliant, unique, nothing-like-it content, but familiar, reliable, plenty-of-things-like-it content. Which, as it happens, is what A.I. is best at.The debate over A.I. in screenwriting is often simplified as, “Could a chatbot write the next ‘Twin Peaks’?” No, at least for now. Nor would anyone necessarily want it to. The bulk of TV production has no interest in generating the next “Twin Peaks” — that is, a wild, confounding creative risk. It is interested in more reboots, more procedurals, more things similar to what you just watched.TV has always relied on formula, not necessarily in a bad way. It iterates, it churns out slight variations on a theme, it provides comfort. That’s what has long made strictly formatted shows like “Law & Order” such reliable, relaxing prime-time companions. That’s also what could make them among the first candidates for A.I. screenwriting.Large language models like ChatGPT work by digesting vast quantities of existing text, identifying patterns and responding to prompts by mimicking what they’ve learned. The more done-to-death a TV idea is, the greater the corpus of text available on it.And, well, there are a lot of “Law & Order” scripts, a lot of superhero plots, a lot of dystopian thrillers. How many writers-contract cycles before you can simply drop the “Harry Potter” novels into the Scriptonator 3000 and let it spit out a multiseason series?In the perceptive words of “Mrs. Davis,” the wildly human comedic thriller about an all-powerful A.I., “Algorithms love clichés.” And there’s a direct line between the unoriginality of the business — things TV critics complain about, like reboots and intellectual-property adaptations and plain old derivative stories — and the ease with which entertainment could become bloated by machine-generated mediocrity.After all, if studios treat writers like machines, asking for more remakes and clones — and if viewers are satisfied with that — it’s easy to imagine the bean counters wanting to skip the middle-human and simply use a program that never dreamed of becoming the next Phoebe Waller-Bridge.And one could reasonably ask, why not? Why not leave the formulas to machines and rely on people only for more innovative work? Beyond the human cost of unemployment, though, there’s an entire ecosystem in which writers come up, often through precisely those workmanlike shows, to learn the ropes.Highly formatted shows like “Law & Order” could be among the early candidates for A.I.-generated scripts. NBCThose same writers may be able to use A.I. tools productively; the WGA is calling for guardrails, not a ban. And the immediate threat of A.I. to writers’ careers may be overstated, as you know if you’ve ever tried to get ChatGPT to tell you a joke. (It’s a big fan of cornball “Why did the …” and “What do you call a …” constructions.) Some speculations, like the director Joe Russo’s musing that A.I. some day might be able to whip up a rom-com starring your avatar and Marilyn Monroe’s, feel like science fiction.But science fiction has a way of becoming science fact. A year ago, ChatGPT wasn’t even available to the public. The last time the writers went on strike, in 2007, one of the sticking points involved streaming media, then a niche business involving things like iTunes downloads. Today, streaming has swallowed the industry.The potential rise of A.I. has workplace implications for writers, but it’s not only a labor issue. We, too, have a stake in the war with the storybots. A culture that is fed entirely by regurgitating existing ideas is a stagnant one. We need invention, experimentation and, yes, failure, in order to advance and evolve. The logical conclusion of an algorithmicized, “more like what you just watched” entertainment industry is a popular culture that just … stops.Maybe someday A.I. will be capable of genuine invention. It’s also possible that what “invention” means for advanced A.I. will be different from anything we’re used to — it might be wondrous or weird or incomprehensible. At that point, there’s a whole discussion we can have about what “creativity” actually means and whether it is by definition limited to humans.But what we do know is that, in this timeline, it is a human skill to create a story that surprises, challenges, frustrates, discovers ideas that did not exist before. Whether we care about that — whether we value it over an unlimited supply of reliable, good-enough menu options — is, for now, still our choice. More