The question is still open

In 1950, Turing made a prediction: by the year 2000, a machine might become good enough that an average questioner would identify it correctly only 70 percent of the time after five minutes of conversation.

That prediction was not clearly fulfilled. Some programs fooled some judges, but small demonstrations were not enough to show that a machine could do it reliably.

Even reliable success would leave another question. Would it show that the machine was thinking, or only that it had become very good at imitation?

His game gave the debate something observable to measure. It did not end the debate.

ELIZA makes the limit easy to see. Behavior can be convincing while the mechanism beneath it is much simpler than it appears.

The original question remains open.

The question is still open

That prediction was not clearly fulfilled. Some programs fooled some judges, but small demonstrations were not enough to show that a machine could do it reliably.

Even reliable success would leave another question. Would it show that the machine was thinking, or only that it had become very good at imitation?

His game gave the debate something observable to measure. It did not end the debate.

ELIZA makes the limit easy to see. Behavior can be convincing while the mechanism beneath it is much simpler than it appears.

The original question remains open.