Intelligence, human and artificial

This afternoon's session that Krista gave on AI was fascinating, both in the material that was presented, and in how the students got thoroughly involved in the discussion. She began by discussing: What is intelligence? Is it possible for machines to be intelligent? How would we know? etc.

Then, we watched various segments of one of my favorite movies, 2001: A Space Odyssey, specifically those involving HAL (which, by the way, according to the book means Heuristically programmed ALgorithmic computer). Krista would show a scene and then ask thought-provoking questions which prompted a discussion.

For example, the first part was of HAL being interviewed for a television show, where he describes that he, as part of the "HAL 9000 series," is perfectly reliable and has never committed an error, and that he "enjoys" working with humans and finds them "stimulating." It amazed me that the students had no qualms at all about the idea that a machine can have emotions; they also made the distinction of whether the emotions were real or simulated for human consumption. I saw lights turn on in their heads when Krista confronted them with the questions: But what are emotions? What does it mean to say they are "real," and how would we know? This pattern continued for the various other segments.

Another scene showed HAL asking to see some of the drawings of a crew member (one of the two humans), only for HAL to segue to more personal questions asking how he (the human) felt about the mission (i.e., HAL's main concern). I really liked Jordi's response to this, that HAL was doing what most humans do, engaging in "small talk" only to then get at what they really have on their minds, making the broader point that it is these kinds of behaviors that are more indicative of whether or not HAL is "intelligent."

The final scene had all the students riveted, where HAL is making every emotional appeal to the human to not disconnect him, and then reverting to a more primitive thought process as the higher-order functions of his "brain" are being disabled, ending in the singing of "Daisy" (which Krista did a good job in explaining its significance).

What I got from this discussion is that this new generation of students is quite open to the eventuality of machine intelligence, and maybe that seemingly human qualities such as emotions may be transferrable to machines.

We then discussed other aspects of AI, such as reasoning, natural language processing, vision, the Turing test, robots, artifical life - lots of great stuff, and very thought provoking. A video was shown of Karl Sims' "virtual creatures," where computer-simulated evolution is used to program behavior such as swimming and walking/jumping. The students were really engaged and fascinated that such behavior could be the result of a computer-driven evolution as opposed to more intentional human-directed programming.

This was the first day where I saw every single student very engaged and contributing to the discussion. Credit to Krista for getting everyone to speak, and for presenting the material in an interesting and stimulating way. She is a natural and gifted teacher. And judging from the students' blogs, I think she is a big hit with the students. Just a few of the responses:

"Today's lecture on AI was extremely stimulating."

"Todays lessons were especially intriguing and i have become very interested in how AI works."

"I enjoyed the lecture in the afternoon. AI interests me because one day I hope computers will be almost as smart as humans."

"I have to say today lesson is really cool. I learned a lot of from Krista about AI, especially the part that we can use our brain to control the robot."

"Overall I enjoyed the session and felt Krista did a good job preparing for it."

"I have to admit that it was the best one so far."

Wow.