LIs artificial intelligence (AI) booming? Are the very high expectations placed on this technology capable of creating text or images likely to be disappointed? Signs point in this direction. Since the overwhelming success of the ChatGPT conversational robot, launched on November 30, 2022, the digital giants have been engaged in a race to dominate this sector, followed by a myriad of start-ups with sometimes stratospheric valuations, such as OpenAI, at $80 billion. (72 billion euros). Tech company executives describe AI as an innovation “deeper than fire or electricity” (Google) or as an unprecedented technological revolution, likely to open an era “abundance” where machines would free humans from much of their work (OpenAI).
“Beware of AI euphoria”, warns the editorialist of Financial Times Rana Foroohar. “Even if you believe that AI is the equivalent of electricity or the Internet, we are only at the beginning of a highly complex transformation that will last several decades and is far from a foregone conclusion”she warns.
“AI is a bubble”, believes digital essayist Cory Doctorow. Asking the question is like wondering “if the pope is Catholic”, abounds the chronicler of Guardian John Naughtonteacher and author of From Gutenberg to Zuckerberg: What You Really Need to Know About the Internet (“from Gutenberg to Zuckerberg: what you really need to know about the Internet”, Quercus, 2012, untranslated).
Turnovers at their lowest
“The hype around AI is reminiscent of the telecom boom and bust of the early 2000s, during the dot-com bubble”writes the financial editorialist of Financial Times, June Yoon. In the five stages of the hype cycle (“craze”) around a technology, theorized by the Gartner instituteAI would be at the second: the “peak of exaggerated expectations”, before “hollow of disillusionment”. The “hype” around AI risks creating disappointments.
“The investment amounts are completely disproportionate, a thousand times too high”, estimates Google engineer François Chollet. AI start-ups would have raised $50 billion in 2023, depending on the site Crunchbase. Sector investments in computer chips and data centers – in order to ensure the computer calculation necessary for training AI models, but also for their operation – are “oversized”notes an investor in the fund Silicon Valley Sequoia : around 50 billion dollars in 2023 and 125 billion in 2025, according to the Bernstein Institute. And 1,000 billion will be spent in four years, even declared Jensen Huang, CEO of chip leader Nvidia.
You have 37.07% of this article left to read. The rest is reserved for subscribers.