This artificial intelligence (IA) is advancing at a frenetic pace. since access ChatGPT Opened to the public last November technology sector accelerated its commitment to bring to market tools that provide machines with increasingly stronger responsiveness. Microsoft He has taken the lead in an “out of control” business race that could bring him billions of dollars in profits, but whose social, economic and cultural implications pose “deep risks to humanity.”
This fear is what prompted more than 1,000 individuals to sign an open letter asking AI labs to “immediately suspend training of more powerful systems for at least 6 months”. GPT-4. This is the latest generation of the language model. turn on AIHe is the creator of ChatGPT, released March 14 and that Microsoft has integrated into its services, from the Bing search engine to programs like Word or Excel. “This pause must be public and verifiable (…) if it cannot be done quickly, governments must intervene and initiate a moratorium,” they explain.
The letter bears the signatures of 1,125 experts from both the business world and especially the academic world. Significantly, the first signatory is Canadian computer scientist Yoshua Bengio, one of the pioneers in the advancement of technology. deep learning, the technology on which systems such as ChatGPT are based. Among the signatories are names such as: Elon MuskCEO of Tesla, Twitter and SpaceX, Steve WozniakApple co-founder, intellectual Yuval Noah Harari or a politician Andrew Yang, Democratic presidential candidate of the United States in 2020. They are also accompanied by co-founders of platforms such as Skype, Pinterest or Ripple.
unreliable artificial intelligence
Although these are the most popular names, most of the signatories are renowned academics and AI researchers from around the world. “This is a technology Carles Sierra, Director of the Artificial Intelligence Research Institute, from the Supreme Council for Scientific Research (CSIC) (IIIA) whose goal is to create an unreliable text that resembles human language but has a very important underlying problem: it cannot guarantee that what it says is true. , to EL PERIÓDICO in a telephone conversation.
This lack of credibility means the results are unpredictable, but potentially just as “dangerous” as their “likes”, according to Sierra. social networks. reason for request arrangement protecting citizens from the “reckless” strategy of tech giants like Microsoft.
Among the prestigious national experts who signed this letter Ramon Lopez de Mantarasdirector of the IIA; Francesc Giralthonorary professor at the University of Rovira i Virgili; anyone Matthew ValeroBarcelona Supercomputing Center – director of the National Supercomputing Center.
Most of the 1,125 people who signed the letter are experts from universities and artificial intelligence research centers. “Scholars can speak more clearly about the limitations that exist in the tech industry,” says Sierra. “The scariest thing is when companies open a multi-billion dollar melon and we already know what happens when money is involved.”
Still, this letter is viewed with skepticism by some experts. Emily M. BenderProfessor of computational linguistics at the University of Washington stated Despite agreeing with some criticism, the wording of this phrase “helps” exaggerate the capabilities of artificial intelligence and contributes to the exaggeration of giants like Microsoft or Google.