Wednesday, March 05, 2014

Androids and Cyborgs and Robots, Oh My!

While listening to one of my favourite radio programs Spark, it occurred to me that even Nora, who is as well-versed in high tech as anyone, was not making clear distinctions between robots, androids, and cyborgs.  Now while that might not appear to make much difference in this crazy world, it has been a decades long study for me, as part of my pursuit to understand this question: "What is a person?"

You might think that that's a simple question. You're a person, I'm a person, most of the readers of this blog are people (other than the spiders and other web-crawlers which scan it looking for email addresses and keywords, but they are clearly not people, right?).  Personhood is a fundamental aspect of our world, one that arises whenever we talk about life and death, whether it is when personhood begins, or when it ends. It comes up because in a world of scarce resources, where people live and die, ending a life by another person's hand can be construed as compassion, or as murder.

In this blog, I'm not going to consider the personhood of you, your family or your friends. As if life isn't already complicated enough, we soon will face the question of non-human personhood, the arising of artificial intelligences both complex and rich enough in perception that we will bestow upon them the title of person.  And I fear that non-human persons are going to be very, very different from us.

But first, the definitions:

Robots are mechanically-produced machines that can accomplish either a narrow or wide range of tasks, but are distinguished by their lack of conciousness, and their need to be programmed to learn their tasks. Even though there are some fantastic examples of making it easier to program robots (see Rodney Brooks' Ted Talk) a robot is what we call a machine that can't think for itself.  A Roomba is nice example of a robot, and the ones that assemble cars, event when they do something much more creative than that.

Androids are humanoid-shaped mechanical (vs. biological) constructs that are machine-reproduced but have an AI core that is sufficiently advanced to pass the Turing test, i.e. another person would think they are a person too. Data of Star Trek:TNG is a prototypical android, as is David in Ridley Scott's Prometheus, and my most recent favourite android, Dorian in the Fox TV show Almost Human (Canadian link) (US Link).

Cyborgs (cybernetic organisms) are at the core, biologically reproduced people who have surgical enhancements that incorporate electrical and mechanical components integrated with their body and nerves. The Borg in Star Trek:TNG represent the cyborg well. Some might say that anyone with artificial limbs might be a cyborg, but I'd reserve that phrase for enhancements that have some electromechanical elements, and that provide the cyborg with greater capability than they had previously. Oscar Pistorius, the so-called Blade Runner, who won gold at the 2012 Paralympics is by this definition more a cyborg than say Terry Fox, since Terry's artificial leg was clearly an impairment compared to his biological leg, while Oscar's foot replacements were an enhancement. By logical extrapolation, and I'm sure this was a consideration in the debate as to whether he should have been allowed to compete, I predict that the Paralympics will grow into a showcase of enhanced prosthetics where the athletes will eventually break many of the records of the comparable Olympic athletes. One could argue that Canadian Steve Mann and anyone wearing Google Glass might be called a cyborg, but I would restrict the definition to people who cannot remove their cybernetic components without surgery, or a large amount of pain.

In my 35 year engineering career, I have always gravitated to the edges, the boundaries of things. From the edge you can see the whole field, and how one thing connects with another. When two things connect at their boundaries it is called an interface, and whether that interface has been between lego pieces, software libraries, or people interacting with software, I like to work at the interface, at the boundary.

The interesting boundary for this blog is the line between a robot and an android. Here's the thing: Don't be fooled by appearance. Many robot designers, especially in Japan where the effort to create a familiar-looking robot is the most active, are working hard to make robots look like androids. It's amazingly successful. It seems to only take a pair of large eyes that can blink to enable some nurturing people to want to take care of it, and only a hint of curves to make some people want to have sex with it. But developing feelings for a robot doesn't imbue it with the capacity to think or feel.

Still, many people will learn to love their home robots and ascribe to them behaviours and feelings that have all to do with the person's projections. That's not necessarily bad. If a home robot can sense agitation, move over to give a person a hug and tell them they're loved, that might be a pretty good robot, and has many good social contributions. But that alone doesn't make it alive, and it isn't doesn't make it a person.

In the next blog, I'll go more deeply into what a robot needs to become an android, and which the Cylons are.