Cite?
Because the data I see shows the opposite:
Dads in particular -
Fathers’ time with children nearly quadrupled – 1965 dads spent a daily average of just 16 minutes with their kids, while today’s fathers spend about 59 minutes a day caring for them.
I think jobs (paid and unpaid) that neurologically require emotional connection will be one of the few jobs that are protected from robotics.
Therapists, parents, childcare, romantic partner, hospice worker, etc. I don’t know if people will ever truly accept these jobs being done by robots, but who knows what the future holds.
What I think will happen with these jobs is that we will have AI input about the emotional states. Like a therapist will have glasses and an earpiece where an AI examines the patients body language, tone of voice, etc and gives the therapist ideas of what is going on in the patients mind.
The same with childcare, maybe AI glasses will give the caregiver information that the child is feeling hurt or hungry, or about to have a meltdown, or feeling ignored, etc.
People will still want these jobs done by a human, but humans will have an AI watching the interaction and always giving the human feedback about the wants, needs, boundaries, hidden thoughts, etc of the child or the patient. Humans will still do the jobs, but AI will make emotional connection better in these positions by reading endless subtle signs that humans mostly miss about the wants, needs, boundaries, etc of the other person.
Right…but I think that only looks appealing in the movies and TV because sexy robot characters are played by real life attractive humans. IRL, at least based on any current or near future technology, even the most “realistic” female robots look like a sort of uncanny animatronic silicone sex doll.
One advantage humanoid robots have in the previously mentioned jobs inside people’s homes is that homes are not going to be retooled to fit just the robot. So giving it a humanoid form means that by default they will be able to reach and walk to whereever humans can already reach.
Sometimes that could even be a disadvantage, for instance, in electrical or plumbing work, being really tiny would help you fit into spaces humans can’t. Even though the vast majority of the work there is already human-accessible.
But as an assistant to every day household chores, as long as humans sometimes do them themselves, robots with humanoid form can take over if the human can’t or doesn’t want to do it. In order to get rid of that, not only would the robots have to be cheap and reliable enough to rarely break down and be affordable by almost everyone, like washing machines, but homes would have to be entirely redesigned around the robots, in order to save space where humans no longer need to go, and homes can last for over 100 years. If you wanted to make a dishwashing or clothes folding robot, if you gave it a humanoid form, then in the vast majority of cases it would be able to do its job out of the box.
Well right off is the fact that quite a few people are forming emotional connections to their AI agents knowing full well that the other side is not sentient. AI therapy is a thing, not ready for prime time, but that hasn’t stopped people from using it in that way, and not always to bad results.
And many of us form strong emotional bonds with our very non-human form four legged friends. Dang I knew kids who got very upset when their damn Tamogoochi got sick.
I don’t think it takes much for humans to project huge amounts onto even a basketball with a face painted on it.
Then again, there is a market for those silicone dolls, and presumably would be even more for one capable of moving and talking on its own. And it is possible, with enough development, to get past the uncanny valley. So, mostly not yet, but maybe eventually?
Of course, we do have dishwashing robots, and they look nothing at all like humans.
Seems like this could be a better approach all around.
Different forms make it clear that the robot and human have different roles:
Parents might prefer the distinction to alleviate the concern that the child becomes attached to the robot.
I’m not sure how much a childcare robot is needed in the home. If you have a generic housework robot, then that handles a lot of the heavy lifting of childcare already. Make dinner, throwaway this diaper, wash these messy clothes, but let me feed, change, and dress the child.
The first applications of human interacting robots will be hospitals, daycares, and retirement homes. These robots could be non-humanoid since they are just augmenting – not replacing – the human caregiver. And again, using a humanoid robot might send the wrong message to the customers. “We’re replacing our nurses with robots, but at least they’ll like human.”
Right, because the first place you want to test your unproven robot technology is handling small, delicate, unpredictable people.
It’s incorrect to say “the first applications of human interacting robots will be” anything. You need past tense, there. And the conclusion of the sentence is “customer service chatbots”.
Ha, right. My phrasing was dumb.
I meant robots taking care of humans will start in those commercial spaces before moving into the home.
You’ll be able to leave your kid with a daycare robot long before a home robot.
Any time a robots had to fit into a place where it is directly replacing a human being, rather than being a construction built on its own. If you want a machine to sit in a truck to directly replace a human being, it’s easiest to simply use a humanoid robot rather that construct a device (as they did on Mythbusters, for instance) to press on the pedals and turn the wheel. A humanoid device could simply sit in the driver’s space, without any need for complex fitting of controls into place. You could use a humanoid robot as Test Pilot in an airplane or other craft, particularly those in dangerous environments (deep sea submersibles, etc.) There’s a reason that we build crash test dummies in the shape of people, after all.
The suggestion of plumbers is a good one – robots expected to do repair or maintenance work in homes where they might be expected to be generalists would find it easier to navigate if they had humanoid form, ppossib;ly being able to plug into attachments. A robot that could move around a house, open panels and unscrew things, then effect repairs with detachable specialist devices would work pretty well.
And, as noted, robots that have to interact with people would be the best case for humanoiud robots. Philip K. Dick, considering the issue, suggested that the reason robots might have to pass for human would be if they had to infiltrate people, a notion that gave us the story second variety, which probably inspired the “flash-forward” scenes in The Terminator and the entire movie Screamers. We could use the same idea directly to make assassin robots, but a better and more humane application would be care-taker robots for children and the elderly, and, as suggested above, sexbots (which you know is going to happen. We already have very humanoid sex dolls. It’s only a short step to sex robots. I think an awful lot of technology has been inspired by horny engineers.)
Our human form is a result of evolution, which is the long term process in which any given life form adjusts to its environment. I can’t think of any job that would require or even benefit from our human form. Robotics are stronger, faster, and virtually tireless. We are the opposite.
Thinking jobs? Well, computer systems have reached the point where not only are they incredibly fast, they are superior in their ability to calculate correctly. Magnus Carlson, considered by many to be the GOAT in super high level chess, views playing engines as virtually hopeless, comparing it to “a bicycle racing against a motorcycle”. Even consumer-grade hardware and mobile phones, he feels, are now powerful enough to defeat the strongest grandmasters in the world.
Maybe if you already have humanoid robots lying around, or if your robot also has to be able to accomplish other tasks. But if I were building a robot just for the purpose of driving an already-existing car, I’d attach a piston to each pedal and to the gearshift and a stepper-motor to the steering wheel, with just enough framework in between to hold all the moving parts in the right places. Not only would this simplify the construction, it’d also simplify the software: There’s a lot of computation that goes into moving the dozen or so actuators involved in a humanoid hand and arm, as opposed to the single actuator for a direct steering-wheel-turner.
That’s pretty much what I had in mind.
Humanoid is good for a general-purpose robot that has to get inti human spaces and do a random collection of tasks (and maybe even interact with people), whereas you Mythbusters-style car driving setup is built for a very specific task – probably even a specific vehicle. But a humanoid robot could drive any car or truck, and fix your sink as well.
The appeal of the humanoid robot to some degree is the appeal of artificial true general intelligence. Humans are in many ways very good at adapting to a wide variety of circumstances and sorts of challenges in their environments, physically and cognitively. Our form and brains are not maximized for any single niche but do pretty well in a wide variety of them instead, and creating a robot that is similarly adaptable, and able to co-exist in our human built environment, is the challenge. Much easier to create a single task thing.
Even if you want a generalist robot, though, I sure know that there are a very large number of tasks for which I’d like to have had an extra hand or two. And the only way that extra hands would ever be a drawback would be if they got in the way, but it wouldn’t be too hard to make them fold away into a gap in the torso, or the like.
I auspect a human on rollerskates is still humanoid-shaped.
Agreed – and I even spoke above about attachments. But my point is that if you want your general-purpose robot to fit into human spaces and be able to do human-like tasks, it would make sense to have a humanoid shape and size., even if it could add or retract extra stuff.
Yeah; for the purposes of this argument “human shaped” is probably a pretty loose requirement. A robot with built in roller skates, retractable extra arms and a radar dish instead of a head would still probably qualify as “humanoid” even though no actual human is built like that.
Maybe not the radar dish. You really need something at least vaguely resembling a human face for people to think it’s humanoid. Just look at all the anthropomorphized non-living objects (cars, toasters, televisions, whatever) in cartoons and comic strips. They pretty much all have two eyes and a mouth; nose optional.