I think that Cecil is missing a very large point when he makes the statements “What’s the point?”, and “It’s creepy”. Neither of these points actually are of any relevance, in the long run. Basically, the reason why you may see robotics emulating humans is mainly to capture human interest (or simply due to the eccentricities of the designer) - the technology is very much in the rough. However, the irrelevancies of those ideas is that, simply, the main reason behind a desire for there to be androids is the simple reason we continue to make machines - we, as humans, are an inherently lazy lot, and anytime a machine can do a job for us, we jump at the chance to let it do so.
So, what does having an android versus a machine mean? It means you essentially have an unquestioning machine capable of reasoning. Essentially, it’s the high tech version of slavery. Lots of sci-fi plots touch upon what happens when things like AIs or robots become self-aware (my favorite is the Mass Effect series), but to the people who essentially have the money to build these machines, they don’t particularly care so much about a sexbot or endearing qualities of the android so much as it’s smart enough to get the job done and inhuman enough to not question its orders.
The scope of this goes far beyond simple “subsets of human functions” minus the bathroom breaks. Already, simple machines are aiding two of the biggest driving forces behind the industry - wealth and power. Having an assembly line of machines is one thing, but given the snowball effect of wealth, power, lack of scruples, and time, it’s not unforeseeable that those seemingly innocuous assembly line workers suddenly become a militia of highly lethal synthesized soldiers.
Course, that’s all speculation. Perhaps mankind will have evolved to the point there they can prevent said things from becoming a societal problem, however, I do simply have to refute the ideas that androids aren’t viable because they’re creepy and pointless as illogical.
I believe what Cecil was saying is that the android form is never the most efficient form for a task. It is much more plausible that and a human you would converse with an intelligent interface and that interface would use multiple specialized machines to accomplish tasks.
Why would a person have a Jetson’s “Rosie” to do the cooking and cleaning when you could have an interface to contact the kitchen machine to make dinner and the laundry machine to clean your clothes?
Of course it is illogical; ‘creepy’ is an emotional response, not a rational argument. You can’t reason people into liking something that have an emotional aversion to. That is why androids, should they ever be perfected, will remain a technological curiosity and never become and everyday part of human life. Humans are illogical, and they don’t like ‘creepy’. Androids are unlikely to overcome a gut reaction.
We already have highly lethal synthesized soldiers. What’s to be gained from making them more humanlike? That would probably mean that they could no longer fly, or quadruple-wield full-auto shotguns, or withstand bullets, or carry a half-ton of ammo, or run at 60 MPH.
Quite a lot in the right situations, actually. Androids designed specifically to emulate human appearance would have an easier time mingling and moving amongst the populace (a common plot mechanic for android-based fiction). Let’s face it, you just can’t quietly slip a mech in the back door of a hit’s target. You also can’t have it rearm itself or commandeer enemy weaponry in the field, pilot enemy vehicles, or pursue someone up a flight of stairs - the human form is well suited to navigating human settlements.
As far as the NEED for this, remember that we live in an age of WMDs - if you go and setup an artillery line outside an armed country you don’t like, the response is going to be their best attempt to one up you by turning your entire country into an irradiated crater. Thus, the answer to those wanting to pull the strings, is not more gun, it’s a quieter, hidden, more accurate gun.
As far as the “why do they need to be humanoid” quesion goes, I think the OP has a point with the laziness argument. A well designed humanoid robot could perform any physical task that a human could, so if I was looking for an all-purpose chore eliminator an android would be ideal. Whereas if I tried to rely on multiple specialized machines, I might have to do certain jobs myself if none of my machines offered those functions. And I might only ever need a particular task done once in my lifetime so it wouldn’t be economical to buy a specialized machine to do it.
Sorry, but I think Cecil’s on the mark with this one. The creep-out factor is pretty high on things like this (I’m reminded of the performance capture version of Beowulf with Anthony Hopkins, which looked a little too creepy for my tastes). If people are weirded out too much by something like this, or an android, then where is the free market incentive to create them in the first place going to come from? People and advertisers are simply not going to shell out money for something that’s going to weird them out too much because it looks too eerily like the real thing, even though they know right away it’s not. Never underestimate the power of emotion (and illogic) in the marketplace. Recall what an expensive lesson Coca Cola learned in 1985 when they switched to New Coke. In blind taste tests, people preferred the taste of New Coke, even though they were complaining about the switch to New Coke. Doesn’t make sense, of course, but then, we’re illogical creatures.
The “uncanny valley” factor seems to overlook something fairly obvious. There is a middle ground for making androids that have human form and functionality but specifically don’t try to look exactly like humans. Try for something more C3PO than Data. It has the humanoid form to use human tools and interfaces because it is a general application tool, has enough of a face to make it easier for humans to relate to, but doesn’t go so far as to fall in uncanny valley.
Roboticists have done research that suggests something like a furby gets postive results when used for facial feedback. So instead of trying to make them look like humans, make them look like teddy bears. Viola - problem solved.
