The conflict between OpenAI and Scarlett Johansson is a good example of how technology companies have acted in recent years.
Summary: OpenAI CEO Sam Altman asked the actress for permission to use her voice in an artificial intelligence (AI) assistant. The idea was to wink at Herthe film directed by Spike Jonze in which Joaquin Phoenix falls in love with the AI voiced by Johansson.
The actress said no.
Even so, OpenAI launched a voice called Sky very similar to Johansson’s. And Altman tweeted “her” during the presentation of the new free version of ChatGPT. Johansson asked for explanations and the company tried to justify itself by saying that she was the voice of another actress (Scarlotte Johanssen, perhaps). But Altman ended up withdrawing Sky and apologizing.
This way of acting is not very different from what we have been suffering for years from technology companies, which continue to remain faithful to the “move fast and break things”, which was Facebook’s internal motto in its early days. years. The engineer Lorena Fernández Álvarez sums it up very well in a tweet: “You throw your hands up that OpenAI has stolen Scarlett Johansson’s voice when she has been doing data extractivism with the entire internet… But progress, if not , does not reach. You know”.
You put your hands up that OpenAI has stolen Scarlett Johansson’s voice when it has been doing data extractivism with the entire internet… But progress, if not, will not come. You know.
— Lorena Fernández Álvarez (@loretahur) May 21, 2024
These companies always present their products with good words, which is quite understandable, of course. For example, in the case of social networks, they promised a better world in which we could contact our friends, share texts and illustrations, discover new ideas… And all this is true.
But it is also true that when someone asks, for example, how the algorithms that select the publications we see work or what companies are going to do with all the information we give them, the platforms usually respond vaguely, like leaving all the decisions behind. in the hands of its users. This usually means that if we have about 17 hours free, we can read the terms of use and modify all the privacy options. Or, why not, believe the discourse that privacy is something out of fashion, very 20th century.
In short, between one thing and another, we have been reading headlines for years about depression in adolescents, political polarization and the destabilization actions of Russian, Chinese and Iranian spies. And all for sharing a couple of memes about Pedro Sánchez.
Something similar is happening with artificial intelligence. These programs are trained with images and texts without companies asking permission or compensating their authors. A bit like what happened to Johansson. With one important difference: with Johansson, OpenAI has preferred to back down. But it would have been more complicated if Altman had made the same case to, what do I know, Escarlata Juánez, an imaginary illustrator from Logroño, as proven by the legal proceedings that creators and media outlets have opened against several AI companies.
It should not be so difficult to have laws that protect Scarlett Johansson’s voice and Escarlata Juánez’s work equally well, as the actress also requests in a statement. Because, otherwise, technology companies will maintain their modus operandi: say good words while the internet fills us with deepfakes, plagiarism and hoaxes… And the response from technology companies will be the same as always: “Is your face yours? And that drawing too? “I don’t know what to tell you, this concern for identity and copyright is something out of fashion, very 20th century.”
I am sure that artificial intelligence will bring interesting and useful projects, just as, for example, I have met very smart people on Twitter (really). But it’s better that these companies take everything a little more calmly, because whenever they go fast, the things they break are ours.
