Seven hundred million data “and then there’s me…”

Generative AI is not a Formula 1 car: it’s a rocket. And sometimes it goes too fast. Even the experts are sometimes overwhelmed…

“Five years ago, we were delighted to differentiate dogs from cats. Today, we have the equivalent of glasses for the visually impaired that describe the scenes in a film very precisely.” In one metaphor, Jean-Baptiste Girardin, Data Product Manager at Malakoff Humanis, has summed up the neural speed at which artificial intelligence (AI) is evolving. “I’d say we have a breakthrough every six months,” says Pierre-Louis Bescond, Data Director at Roquette.

Jean-Baptiste Girardin remembers “a very brilliant PhD student”: “every time he took off in one direction, an article came out shortly afterwards and obtained very good results.” Higher, too fast, more powerful: AI provokes two concomitant fears: the fear of missing out (FOMO) and therefore the fear of becoming obsolete (FOBO), says Eneric Lopez (see elsewhere), director of Microsoft’s National Initiative for AI.

But it’s not only employees who are scared. Organisations themselves are swamped. “There are more and more barriers at the outset when it come to innovating,” says Jean-Baptiste Girardin. “The big AI players, like Meta, Google and OpenAI, are delivering models that set such high performance requirements that…” companies are finding it hard to stand out. “Even our most experienced data scientists are sometimes astounded by the speed at which OpenAI moves.”