Scientists from the Washington Institute conducted a survey of scientists involved in artificial intelligence research. It was aimed at finding out what opinion scientists hold about the impact of AI on society as a whole.
The results showed that 36%, that is, a third of respondents, believe that the development of artificial intelligence can lead to catastrophic consequences, such as a nuclear war or some other catastrophe of a comparable scale.
This will happen within the next century. The researchers’ work is published in the arXiv preprint repository.
At the same time, 73% of scientists believe that artificial intelligence technologies will accelerate the global industrial revolution, which will inevitably be followed by serious upheavals in society.
The development of artificial intelligence, as noted in the paper, is the subject of long discussions. There are radically opposite opinions about whether it is necessary to strive for this and what will happen, what limits in development are acceptable and what can happen if AI reaches these limits.
Other findings included that 74 percent of AI researchers believe that the private sector has too much influence on the field, and that 60 percent believe that the carbon footprint of training large models must be a very serious problem.
In their opinion, the development of artificial intelligence leads to the destruction of the environment.