2023-04-21 00:30:31
On March 29, a group of prominent figures, including from Silicon Valley, published a letter in which they called for the development of artificial intelligence to be halted for at least six months, and to use that time to develop security protocols. We talk about it with Wim Naudé, AI expert at the OECD.
What are the authors afraid of?
“The authors fear GPT4, the latest version of a language model based on artificial intelligence (AI). It is unclear how this model was trained. The fear is that the model can develop itself so quickly that it will become smarter than humans. Then the end could be near for humanity, according to these scientists. For example, AI can see humanity as a threat and declare war on them. Or the threat runs through a AI lock-in, with AI taking over government and passing laws worldwide that can never be changed. There is also a situation conceivable out The Matrixwhere the AI provides incentives, and can thus influence human behavior.”
Are the concerns justified in your eyes?
“Philosophically, yes. But this call also seems to have been prompted by the disappointing results of AI. It’s not the first open letter from this group of Silicon Valley celebrities. In 2015 they published on the website of Future of life an open letter, in which they advocated more research into AI. They thought that the benefits of AI would be immense, and that it could, for example, solve global poverty. Now that this seems to be disappointing eight years later, the benefits apparently no longer outweigh the risks, according to them.
But you could also see the call as a publicity stunt: the tech industry is under great pressure. On the one hand, the authors try to indicate how powerful their creation is – by warning against it. On the other hand, they want to show how virtuous they are with the call.
However, I myself do not think that AI poses an existential risk – no civilization has yet been destroyed by its own technology.”
So we should continue researching AI?
“We especially need to do more research into human intelligence. We can learn a lot from it, because the human brain achieves excellent results compared to AI – while using a fraction of the energy. But we still understand little about it – human consciousness, for example, is still a black box.
We must also continue to research AI, but the tech industry should focus more on the major breakthroughs. The AI gadgets are distracting us from the fundamental technology we are looking for – and the search for which will take longer. For example, according to a growing chorus of commentators, fundamental physics, which provided virtually all of the technologies that underpinned earlier industrial revolutions, is stagnating.”
Could AI Drive Productivity Growth?
“That is also disappointing for the time being. In the United States, only 2 percent of companies use AI. Many companies find AI too expensive and the benefits too limited, and therefore not profitable. In any case, it is not certain that AI will increase productivity, because the application can be very complex and time-consuming.”
Should we worry about our job becoming obsolete?
“No, I do not think so. Some jobs will become redundant, but I don’t see any problems with that. There are more jobs than ever. My kids now have no idea what to choose from the 15,000 or so jobs out there. What type of career we want ourselves is now central. I have also occasionally thought whether I want to continue doing my current job for the rest of my life. And in this generation, people can happily switch careers.”
Literature
Bengio, Y., S. Russell, E. Musk et al. (2023) Pause giant AI experiments: an open letter. Open letter, March 22. Available at futureoflife.org.
#Results #artificial #intelligence #disappointing