Discussion on:

Artificial Intelligence: What happened to the hunt for thinking machines?

The Turing Test is based on the concept of loading a device with a lot of data and then developing algorithms for it to use this data. This is the way technology has been perceived to be used.

During the past fifty years, we have witness a transformation from data processing to information processing. The difference is in that the technology is not just simply processing data, but it is collecting data and then providing information.

This is all fine and dandy for answering questions, but does provide for a conversation.

To have a conversation requires more than data or information processing, it requires intent.

Where is this conversation going? What do I want to say? Why am I having this conversation? What and why did the other person say this?

The "hunt" for a "thinking machine" does not rest with "thinking" but with machine logic that focuses on what the machine can do and why it should do it, if it at all. This, in my opinion, will come more from the robotics than data driven analytics.

I have heard that while no computer has passed the Turing test there have been people who failed it.

But more seriously, the test is based on the assumption that if an entity is intellegent it will behave just like human. This reminds me of an XKCD cartoon where the ants decide there is no intellegent life other than them because the humans have not picked up on their scent trails.

Perhaps Marshall McLuhan's observation that the message depends on the medium also applies to intellegence. Humans have one form, based on our evolution. Dolphins have another form. Computers will eventually have yet another form. I think we are seeing increasingly smart machines, but they are designed to solve problems and not just act like humans.

Just wondering why the author did not mention SIRI... I know she doesn't hold up a great conversation, but definitely has some interesting replies. And why was Apple not mentioned as one of those groups that might be interested in competing? They do seem to be the Tech company most interested in making the human interface more palatable - a conversation based interface would seem to be such, so why not?

"While sci-fi offers artificial intelligences that rival our own, the fiction bears little resemblance to real world AI"

It would be more appropriate to say that so-called real-world 'AI' bears little resemblance to intelligence. Calling "specialized software that helps fly planes and run factory production lines" AI is a tacit admission of failure, at least for now.

I do not mean to imply that the Turing test is the only or best test of AI, but any such test means little unless it requires that actual intelligence is displayed, just as any test of an airplane should require actual flight.

It may be that we are the robots someone else invented!

Wouldn't the best robot would be a biologic ... didn't the Summerians describe this whole story 5000 years ago?

Now if we did it in the last 100 years and it was that good how would we know?

I suspect that modeling minds instead of conversations is more useful for purposes of trying to reach the goal of reproducing "the complexity of human thought in a digital form." To this end, I've constructed a simple model of a cognitive agent that engages in conversations but, unlike chatbots, is specifically designed NOT to fool anyone. (See ProtoThinker.com)