Stork: What would you have said if you had won?
Clarke: I've no idea what my Oscar speech said. I must try and dig it out one day. It must be buried in the Clarkhives.

Stork: Can we turn now to HAL and vision. HAL we see primarily as this big red eye. Where did that red eye come from?
Clarke: HAL's famous red eye was a product of the art department, and I believe that somebody somewhere has the original red eye which they picked up at an auction sale. It doesn't really make much sense because it would really want binocular vision. Why it was red. It gives it an ominous sort of feeling, but it was probably not very practical.

Stork: But, we have the sense that he's omnipresent, that no matter where you go, there's his eye.
Clarke: Big HAL is watching you.

Stork: So you were influenced by big brother?
Clarke: We're entering a world where HAL-type computers are watching us. They're installing television cameras in cities, particularly in high crime areas, which has caused a great deal of consternation but I'm all in favor of it.

Stork: So HAL recognized the drawings of the crewman. He did lip-reading and speech-reading. Tell us about HAL's vision and visual abilities.
Clarke- The one ability that I was doubtful about, and this was Stanley's idea, was his power of lip-reading. First of all, I didnÕt think it was possible for a computer to lip-read. Secondly, why should they bother to give him that facility anyway? ThatÕs an interesting point that I don't think is ever explained. Anyway, it certainly produced the most effective sequence. And, of course, itÕs led to a number of careers in the business."

Stork: But perhaps lip-reading just emerges. Once you've got speech-recognition, once you've got vision. HAL would somehow put them together. After all, you were never taught how to lip-read, never the less you do lip-read.
Clarke: Well, I can't lip-read, except of course I probably do it unconsciously. And now that I'm getting quite deaf, I might be doing more and more lip-reading. But it seems incredible to me that a really good lip-reader can fool people around into thinking that they have normal hearing.

Stork: What kind of problems would HAL have to have solved with these kind of vision problems? He'd have to recognize the crew. How's HAL going to do that?
Clarke: The technical problem in enabling HAL to recognize what's going on around him, the millions and billions of bytes of information a second. What is important, what is irrelevant. ThatÕs a fantastic programming problem, but IÕm sure it could be solved, though I wouldnÕt be able to do it.

Stork: Why is it that humans are so good at it and machines are so bad?
Clarke: I think humans are good at this extracting important images from their surroundings, because if we didnÕt know that that sabre-tooth tiger was over there, we wouldn't be here now. So we had been selected for our ability to do just this sort of thing.

Stork: But how is it that HAL can recognize where the crew are. It seems miraculous in some ways, that amidst all this buzzing confusion of sensory input, HAL somehow figures out that what to look at, whom to listen to and so forth.
Clarke: HAL's most important job was to know where the crew was. The job of teaching him what was important was a tremendous one. ItÕs the unexpected. The things that no one could have possibly imagined that cause real problems. I mean, really unbelievable things that weÕve seen examples of in technology of accidents that no one could have dreamed would have ever occurred, and they do.

Stork: So humans are very flexible, and many people feel that computers only do what you tell them, and hence will never be able to anticipate these unexpected events.
Clarke: It's a fallacy to imagine that computers will only do what we tell them to do, because we can extend their abilities. We can tell them to be curious, and they could go on to learn things which we didn't know. That has been done, in fact, in mathematics.

Stork: Do you want to discuss that? I mean work by Doug Lenat on discovering new theorems in mathematics by computers?
Clarke: I don't know much about it.

Stork: So we humans rely on vision a great deal. Much of our world is visual, much of our brain is visual. Do you think you could have a HAL without vision?
Clarke: HAL without vision would be very limited, but there are all sorts of other senses that for instance, the ability to sense electro-magnetic waves. Sound, of course, it could build up a picture by sound as dolphins do. So, although vision is important, itÕs not perhaps all important.