Geoffrey Hinton, a researcher dubbed “the godfather of AI” for his enormous contributions to the field, has resigned from Google out of dissatisfaction with the direction artificial intelligence tools are taking.
In an interview with the New York Times, Hinton warns that it will be hard to prevent the tools he helped create over the past decade from being misused. “I console myself with the usual excuse: if I hadn’t done it, someone else would have done it,” he explains in the interview.
Hinton, who won the prestigious Turing Award in 2018 for his contributions to the field of artificial intelligence, is the developer of many of the techniques that make it possible to create neural networks -a computational model that simulates the behavior of neurons in living beings- capable of to distinguish concrete objects.
His work has made it possible for computers in recent decades to be able to understand with a high level of confidence what happens or what elements are in a given image or audio fragment, a task that until not long ago was considered almost impossible.
In artificial intelligence, these techniques developed by Hinton are known as “deep learning” techniques, and allow algorithms to be trained to detect a cat or dog in an image, for example, using hundreds of thousands of examples. After being exposed to all these images, the system itself ends up inferring the properties that define the essence of the animal in question and is capable of detecting them in new photographs, even if it has not seen them before.
It is a key advance for modern artificial intelligence tools, including those known as generative artificial intelligences, capable of creating images or text from natural language descriptions.
The researcher joined Google in 2013 after the web giant bought the company DNNresearch, which Hinton founded with some of his students at the University of Toronto.
So far, the researcher was happy with the progress and the way Google was developing its AI tools, but he says something changed after OpenAI’s announcement of ChatGPT late last year.
In just a few months, the race to create more and more advanced AI tools has intensified and Hinton has decided to leave the company so that he can talk about the dangers posed by AI without affecting the company, which he believes is acting ” Responsibly”.
Hinton also believes that we are very close to creating generalist artificial intelligences that are much smarter than humans, a scenario that until recently he considered to be a long way off.
His voice thus joins that of other academics and researchers who are beginning to show concern about the rapidity with which artificial intelligence tools are being refined and the fierce competition that has arisen between large technology companies for increasingly complex applications. and with a great social impact.
Last March, several of them published an open letter asking that development in artificial intelligence be stopped for at least six months in order to agree on common guidelines or a regulatory framework that allows progress with fewer risks for the population.
According to the criteria of The Trust Project
