Man likes making predictions about the future. Scientists are human beings, therefore they
make predictions about the progress to be expected in various fields of
research during the coming years, decades and even centuries. These predictions
are widely publicized by the media.
Are scientific
predictions more likely to be satisfied than other predictions of the future?
We might think so, since science is the most rational branch of human
knowledge. What should we do to confirm or disconfirm this surmise? We should apply
the scientific method to the predictions, i.e. wait until the scheduled time has
come and check whether the predictions were fulfilled or not. Such studies are
not usually done. Everyone is prepared to predict or to listen to predictions, but
few bother to check if those anticipations actually came to happen.
There are a few egregious
cases that many people remember. In 1956, the Dartmouth Summer Research Project
on Artificial Intelligence, where the term Artificial Intelligence was coined, predicted
that in less than ten years we would have computer programs capable of beating the
world chess champion, or seamlessly translating between any two human languages.
The total failure of this prediction is obvious: the first target came true 41
years later rather than 10, while the second has not been achieved after almost
60 years. As a result of this failure, research in artificial intelligence
stopped for more than a decade and was not revived until expert systems
reawakened interest in the discipline.
