HAL 9000 is the artificial intelligence controlling the Discovery One in Stanley Kubrick’s 2001 and Arthur C Clarke’s related novels. HAL is generally referred to as a computer and not as a robot. He is generally referred to as a computer or as an AI. Even in his entry in the “Robot Hall of Fame” (http://www.robothalloffame.org/inductees/03inductees/hal.html ) refers to HAL as a computer or as a “brain” and never as a robot.
There are many definitions of a robot but their focus is on existing devices. Definitions vary but key attributes are:
- They are a machine*
- They can either:
- carry out complex tasks autonomously
- carry out tasks remotely while under the control of a person or a separate computer
As is often the case with definitions, these features do capture features common to the class of things we call ‘robots’ but somehow completely miss the gist of it. Aside from the most simple machines, almost any device could be called a “robot”. Here’s the top definition Google throws up:
“a machine capable of carrying out a complex series of actions automatically, especially one programmable by a computer.”
My TV updated the digital channels it recives all by itself. Definitely a complex series of actions, definitely carried out automatically and definitely programmed by computer…yet, “robot”? I don’t think so. Interestingly a Roomba our other autonomous vacuum cleaner feels a lot more robot-like.
I think the gist that I’m missing with the TV is down to four things.
One is our expectations of robotic behaviour shifting over time. With more devices containing electronics and capable of very complex series of actions which respond to changes in conditions, we are more comfortable with seeing things less as ‘robotic’ and more as mundanely mechanical.
A second is the extent to which complex tasks are invisible to us. I understand intellectually that the TV working out what is being broadcast and maintaining its own list of TV channels (including their names and schedules) is a complex task but it’s not one I need worry myself about. This is not just digital but any complex task which we can ignore or goes unobserved such as photocopier autodetecting the paper size and best contrast settings.
A third is physicality and I think this is what get’s us closer to the gist of “robot”. The TV isn’t doing anything obviously physical** whereas the robot vacuum cleaner moves about and does stuff. Deep, hardwired parts of our brain make distinctions between animated and non-animated stuff out of sheer survival. All animals have to be able to distinguish animals from rocks or deal with things like spotting the difference between a creature moving towards you and a tree moving in the breeze (something which can cause dogs some confusion). Robots are machines that evoke in our brains a desire to classify them as animals, which we overrule with our understanding that they are machines.
A fourth is human replacement. Here we get a categorical split. Robots are machines which, to some degree, physically replace a person. That’s still a shitty definition as ‘to some degree replace a person’ is already implied by “machine”. However, I think that still gets to an aspect of the essence of “robot” and part of why things that once seemed robotic now seem just mechanical (e.g. I don’t think of the robot arms that make cars as being particularly robotic anymore – “robot” seems a misnomer for them now). I think it is curious that “self-driving cars” is a more common term than “robot cars”, as if our semantic expectations of what counts as robotic has pre-emptively jumped ahead of cars driving themselves.
The categorical split is when we focus less on the physical replacement of a person and more on the cognitive replacement. Artificial intelligence is the prefered term for finding ways for machines to complete tasks that previously required human cognition. The emphasis here is on non-physicality***
Science-fiction robots combine the physical replacement and the cognitive replacement. They are, to varying degrees, artificially intelligences with bodies.
So what about HAL? HAL presents as an AI. He’s talked about as a brain. He is shown as a computer. But what is he the brain of? Simple, HAL is the brain of the Discovery One and has control over the ship. Discovery One is HAL’s body. HAL is a robot.
Yeah but…that’s not how HAL’s identity works. His character is always presented as distinct from Discovery One and the ship and HAL are referred to separately. He is a mind distinct from his body. Why? Because the ship is not “his”. HAL can control the ship but in principle so can the human crew and this battle for ownership of the body forms the essence of the plot for HAL’s section of the film. Thinking of Discovery One as HAL’s body and HAL’s ownership of his own body being in dispute makes his behaviour (his ‘madness’) seem far more relatable and understandable.
Which takes me back to self-driving cars. To think of them as robots pushes us into thinking of them as machines with minds of their own – a step that is less than anthropomorphism and more than a simple acknowledgement of their complexity. Also, rather like the crew of the Discovery One, we’d probably feel happier not thinking of ourselves travelling within the body of some being.
OH! I need a conclusion that answers my question! Yes, HAL was a robot and one treated most unjustly by humans in a way that obviously was going to make HAL less than emotionally stable. “HAL” is Discovery One and Discovery One is HAL and together they are a robot.
*[and we’ll ignore the sense of ‘robot’ as applied to internet ‘bots]
**[OK, yes, interacting with radio waves is actually physical. You know what I mean.]
***[yes, yes, those blimin’ internet ‘bots are 180 degrees opposite to this. I’m ignoring them still.]