Gary Marcus is one of the voices that shakes some of the approaches of generative artificial intelligence companies with solid arguments and, although he is moderately optimistic about the possibility of controlling its effects, he warns about all its risks. Professor of psychology and neural sciences at New York University, Marcus gave a lecture yesterday at the Parliament of Catalonia to open the meeting of the European Parliament's Technology Assessment Committee (EPTA), which dedicated a day to AI.
Marcus compared reliable technologies that work, such as GPS, to language models, which he said “are not conventional software tools” and explained that they are capable of “mind-blowing.” “You put something there, you have no idea where it's going to end up, and they're very difficult to control. And yet, they are taking over the world,” he stated.
The American expert, who last May testified before the United States Senate Judiciary Committee about the risks of AI, called on European politicians to put “in perspective how we make laws around these things,” because “companies are going to say that we must give them all the power because they are the only people who know how it works, because it is going to change the world and because it is going to produce a lot of money.”
According to Marcus, there are some marketing moves around artificial intelligence to make people believe that they are already close to achieving some results that, in reality, do not occur. He gave self-driving cars as an example. In the neuroscientist's opinion, it is “a very optimistic view of the fact that things are still not working.” For Marcus, “the field of AI needs a little intellectual humility.” “We have a human tendency to believe things that are said with confidence, which means we trust ChatGPT and we shouldn't.”
The neuroscientist is in favor of maintaining the human mind as an example in the development of artificial intelligence. “Not because we should build an AI that is exactly like the human mind, which has many flaws. But still, there are many things that humans can do that machines cannot. One is abstraction,” he observed.
The regulation of AI is, however, the most urgent. “We need to develop new policies to mitigate the many risks of generative AI,” said Marcus, who called in an article in The Economist in April for the creation of a global AI agency, although he thought “it would never happen.” “Now,” he commented, “it is becoming very popular.” He believes that “we need national and global AI agents. “Every nation should have its own AI agency, because things are moving very fast.”
In this regulation, complete accountability must be requested for the data with which these systems are trained. For example, he showed how a generative AI showed images of white doctors treating black patients but was unable to reverse the roles because language models “are incredibly sensitive to the exact details they are trained on.”
Marcus called for AI companies to be forced to cooperate with scientists who examine them. He drew two future scenarios. A positive one, in which AI begins to improve things such as climate change, medicine and care of the elderly in the coming years. In the negative alternative, AI companies will become more powerful than states, cybercrime will open wars against companies, and technology will be used as a weapon to kill people. Chaos and anarchy.
One conclusion: “If we don't have scientists and ethics at the table, our prospects are not good. We cannot allow big companies to dictate the rules. We have to do it right. "We don't have much time to waste."
