Henry Kissinger’s concern about artificial intelligence is valid, as described in a recent David Ignatius column. ‘Artificial Intelligence’ is misleading.
The output from computers is real but it may or may not be intelligent.
Computers cannot program themselves (yet) so AI computing depends upon line-by-line instructions and algorithms devised by humans. These algorithms sift through and interpolate between trillions of bytes of data to produce an “answer.”
These algorithms are very sophisticated and intricate but they are none-the-less human inventions. If the output were the solution to a mathematical problem it would be immediately obvious whether or not it was correct. But if the output was concerned with human behavior or some other non-mathematical entity, it would be impossible to assess its veracity. And there lies the danger. For example, should we rely on the computers to assess the world political situation and following their recommendation to fire nuclear missiles? I hope not.
Police departments in many cities tried to use AI to isolate areas of crime and profile individuals and it was a total failure. After wasting millions on it the system it was scrapped. (This was the subject of a recent PBS program).
So let’s beware of AI and not follow its lure.
Len St. Clare