Can the Chinese Room Think?

In 1980, a philosopher named John Searle came up with a famous thought experiment called the Chinese Room experiment. It goes like this:

“Searle imagines himself alone in a room following a computer program for responding to Chinese characters slipped under the door. Searle understands nothing of Chinese, and yet, by following the program for manipulating symbols and numerals just as a computer does, he sends appropriate strings of Chinese characters back out under the door, and this leads those outside to mistakenly suppose there is a Chinese speaker in the room.”

 

The narrow conclusion was that a digital computer could only appear to understand language, but it could never truly understand it. The broader conclusion was that the thought experiment disproved the idea that the mind was just a computing/analytical machine. Its implications thus fell into semantics, language, mind, consciousness, computer science and cognitive science.

 

Does Google’s Pathways Language Model (PaLM) prove that the Chinese Room can think? But first, what is Pathways?

“We’d like to train one (AI) model that can not only handle many separate tasks, but also draw upon and combine its existing skills to learn new tasks faster and more effectively. That way what a model learns by training on one task – say, learning how aerial images can predict the elevation of a landscape – could help it learn another task — say, predicting how flood waters will flow through that terrain.”

 

Here’s one example of the level of “understanding” PaLM can do.

Question:
Roger has 5 tennis balls. He buys 2 more cans of tennis balls. Each can has 3 tennis balls. How many tennis balls does he have now?

Answer:
The answer is 11.

 

But wait, PaLM can even “explain” how it arrived at the answer.

Input:
Jennifer looked out her window and sees a really cool cloud below her. She unbuckles her seatbelt and heads to the bathroom. Is Jennifer probably traveling more than 300 miles per hour relative to the earth?

Model Output:
300 miles per hour is about 480 km/h. This is about the speed of a commercial airplane. Clouds are usually below airplanes, so Jennifer is probably on an airplane.

The answer is “yes”.”

 

Many such examples certainly suggest that the computer is reasoning. All of which is why Alex Tabarrok writes that:

“Searle’s thought experiment was first posed at a time when the output from AI looked stilted, limited, mechanical. It was easy to imagine that there was a difference in kind. Now the output from AI looks fluid, general, human. It’s harder to imagine there is a difference in kind. The sheer ability of AI to reason, counter-balances our initial intuition, bias and hubris.”