Report comment

AI, currently, is not intelligence, it is programmatic problem solving, in a very limited sphere. Those that are involved in true AI research will tell you that it is 50-100 years away, and they become frustrated at media and other commentators indicating it is just around the corner. See in particular the work of Prof Murray Shanahan, Professor of Cognitive Robotics, Imperial College London.

Currently an “AI” can play Chess very well, Deep Blue as an example beating a world champion in match conditions. “AI” can also play Go, a more complex game than Chess, with Google’s DeepMind beating a world champion in match conditions.

But Deep Blue cannot play Go. DeepMind Cannot play Chess.

We will not have true AI, as opposed to problem solving, until a device such as Deep Blue designed to play Chess, learns to play Go, and does so well. That is true Intelligence, and why AI is so far off.

Lawyers are seldom presented with a problem in isolation, current state “AI” may provide some assisted decision support but it is decades away from replacing what lawyers do well.

The real problem is that key decision makers, and the public, believe the hype that is pushed at them and make bad decisions based on misinformation.