Geoffrey Hinton and Yoshua Bengio
Posted: Thu Feb 13, 2025 10:21 am
The researcher, whose work on Deep Learning paved the way for “Large Language Models” such as ChatGPT, is worried. But not all scientists share his vision of the future . He points out several dangers : seeing the machine surpass its creator; job cuts accelerate ( IBM has declared that it is pausing 7,800 recruitments); “killer robots” become a reality; or even disinformation threaten democracy. His words echo those of the open letter that called, last March, for a “pause” in AI research because of the “risks for society and humanity.” But, “ it is not because a machine will be superintelligent that it will automatically want to dominate humanity ” reacted Yann LeCun .
The winner of the Turing Award in 2019 (alongside…), Chief ig database AI Scientist of Meta, criticizes an “ obscurantist ” vision. He remains convinced that AI “can lead to a renaissance of humanity, a new century of Enlightenment ” . Beyond the simplistic oppositions ( that technical progress has always aroused ) , a third way is possible , according to researchers at the Distributed AI Research Institute . The urgency. To look into the “exploitation of workers” ( the first union was created on Monday in Kenya ) that drive AI, the “data plundering” that feeds these systems, the “explosion of false content”, and to slow down the “concentration of power” .
Yoshua Bengio affirms it: “We must slow down and regulate!” The AI Act is already under discussion in Brussels . Across the Atlantic, the White House brought together Sam Altman (OpenAI), Satya Nadella (Microsoft), Sundar Pichai (Google) and Dario Amodei (Anthropic) on Thursday, around Joe Biden, for a "frank" discussion on the risks linked to technology .No artificial intelligence without collective intelligence. We will have to trust in progress, on the model of the old dream of the early days of the Web, which we have just celebrated its 30th anniversary , by advocating an "open source" AI system.
The winner of the Turing Award in 2019 (alongside…), Chief ig database AI Scientist of Meta, criticizes an “ obscurantist ” vision. He remains convinced that AI “can lead to a renaissance of humanity, a new century of Enlightenment ” . Beyond the simplistic oppositions ( that technical progress has always aroused ) , a third way is possible , according to researchers at the Distributed AI Research Institute . The urgency. To look into the “exploitation of workers” ( the first union was created on Monday in Kenya ) that drive AI, the “data plundering” that feeds these systems, the “explosion of false content”, and to slow down the “concentration of power” .
Yoshua Bengio affirms it: “We must slow down and regulate!” The AI Act is already under discussion in Brussels . Across the Atlantic, the White House brought together Sam Altman (OpenAI), Satya Nadella (Microsoft), Sundar Pichai (Google) and Dario Amodei (Anthropic) on Thursday, around Joe Biden, for a "frank" discussion on the risks linked to technology .No artificial intelligence without collective intelligence. We will have to trust in progress, on the model of the old dream of the early days of the Web, which we have just celebrated its 30th anniversary , by advocating an "open source" AI system.