Set in the near future after there is a breakthrough in A.I. research, these robots are able to engage in a conversation with us and even provide consultation in many different areas ranging from medical advice to choosing a soulmate. Despite the fact that all these robots are hardwired to avoid load capacity exceeding 100kg for 60 seconds, they all come with powerful heavy duty motors that can allow a single robot to lift up a 30 ton fire engine for a full minute! What could be a good reason for introducing these talking "forklifts" into our society?
Because we dont have production lines that build them weaker
We don't need a lot of processing power for something like a greeting card that sings when you open it, yet the chips inside them have been steadily getting better and better. They are basically over engineered for their purpose. The reason? The amount of low-end chips we need on a daily basis is so small that it's just not economical to keep building them, besides improvements on energy efficiency and material savings. With the production capacity for low-end stuff being scaled down and shut down indefinitely, you have to look for the next low-end chip in line to put in.
The components that build your robot are simply the lowest end and most economic. They don't build smaller actuators on economically viable scales to power the arms, legs and torso, and the smaller ones they do have need adaptations to bridge the length of an arm etc. So the manufacturer simply picks the cheapest functional actuator, which with current tech simply means arms and legs capable of supporting 30 tons of weight.
The "Robots and Public Safety" law of two thousand and hurfty nurf was enacted following the Great Earthquake in [insert precarious metropolis here] when it was clear that the scale of the disaster would have been much reduced if there was a greater "heavy lift" capacity available to help clear rubble and search for survivors. A wide variety of humanoid robots were already in place, but their intrinsically safe motor systems limited their usefulness in this role.
By having an "emergency rescue" mode that can be activated in the right circumstances, the army of regular service robots can be pressed into service at a moment's notice to deal with this kind of situation much more quickly. The usefulness of this capability was obvious, and similar rulings were enacted across the world.
Cities and countries in less tectonically exciting parts of the world still benefit from the emergency system when dealing with smaller scale everyday accidents and emergencies. Whilst some claim that it would make more sense to have a dedicated store of emergency robots in rapid deployment silos and not have your housekeeper be able to rip you in half, it was pointed out that deployment times might possibly maybe be too low, and anyway that sounded like they just wanted all that juicy public safety money to themselves and having sexbots that can pull people out of lahars was totally worth the considerable increase in cost of what had been quite affordable bits of machinery. After all, if it saves the life of just one child, all the risks and costs are justified. You don't want the children to die, do you?
For the same reason cars can reach 150km/h even though you pretty much never need to drive that fast. Or why cars can carry far more than their typical payload and most people will never load their car up anywhere near the limit. Or why some people have SUVs even though they don't need an SUV 99% of the time. They have it just in case they need it or for that one time every year, or every few years they need it.
I'm sure it helps on moving day. Invent one task that most people have to do at some point. A lot of furniture in the near-future could be made of metal instead of wood or plastic. Or perhaps it is standard practice that when moving you load all your stuff into something like a shipping container and then need to maneuver it, or re-position it. Maybe parallel parking is a pain and it's easier just to get the robot to move the car into the parking spot for you.
Hell, it could even be that the robots come with their own "maintenance pod" that is really heavy so the robot is just made to be able to move it on its own. Otherwise whenever you needed to move it when re-arranging your furniture, it would have be a huge operation where you might need to get specialized equipment and workers.
People refuse to let AIs drive cars.
While it would be possible to let AIs drive cars, people don't want to let them, and have pushed laws to force allowance of humans driving cars. Humans being humans, humans often get truck or sleepy and drive cars and hit people.
With modern medical technology and science, cars are the main cause of human death. The robots are designed to be able to stop even very heavy cars or trucks, to negate the need for people to not drive cars. It was deemed cheaper to supply everyone with super strong robots than to defeat the human driving lobby.
Since, as you explained in the comments, this robots put work even into holding a weight against gravity, they resemble humans muscles in their way of working: we get tired after holding a weight for some time.
This means that, like in human muscles, these androids also do a partial activation, depending on the needed force: we can jump 1 meter high or more with our leg, but we don't always put full strength when we use them: we can gently move away a puppy sleeping close to our feet without having it fly away.
Using a single type of muscle at lower power has the benefit of reducing costs, since it allows for mass production.
It's not as hard as you think.
Robots with rechargeable batteries have an unlimited supply of energy, but their design limits their power output, which is the amount of energy expended per time. Lifting an object requires a fixed amount of energy, but arbitrarily little power given sufficient time. Even a human can lift a 30-ton fire truck given the appropriate jack and a bit of time. A robot merely needs to lay under the fire truck and jack it up one millimeter at a time over the course of several minutes/hours, and can do so with even a very low-wattage motor.
You don't need to be powerful to lift heavy weights, you just need mechanical advantage and time. The paltry 10-Watt motor from an electric razor is sufficient to lift a fire truck by 10cm in about an hour. Once lifted, keeping it there requires no energy whatsoever. These robots aren't designed with large motors, but have gears and levers that allow even tasks with huge energy requirements to be accomplished at very low power.
Just call it materials technology. Today, AI on supercomputers actually IS working on super materials, doing quantum chemistry analysis of combinations of molecules to determine characteristics. This computation, although expensive and time consuming, is actually cheaper than trying to physically synthesize compounds; we want some confirmation we will be successful before we try to invent a chemical protocol to actually synthesize such materials.
So, in the future, say we have discovered such things, super steel and super magnets and such, and they are cheap, and of course we build robots with them. Few manufacturers build with anything else.
Kind of like steel or aluminum today. Aluminum was once the most expensive metal in the world, ounce for ounce more expensive than gold. Now it's so cheap we make packages designed to be thrown in the garbage out of aluminum (like soda cans), because it is so useful.
So in the future we have a new products like that, new metals you get to name, new magnets and motors.
So the robots, like everything else, are build of these, and they get this super lifting power for free. They aren't "built" for it, it is just a side effect of being built with the same ubiquitous quantum engineered materials as every other machine in the world. Nothing breaks or bends or cracks or deforms unless it is specifically designed to do so.
humanoid robots designed solely for social purposes
They are "big" friendly robots. They are big to be able to lift 30 tons.
They will lift you up to talk to you:
They will lift your car up to talk to you when you are in your car:
They will eat your car: (OK, kidding, maybe just nibble 🙂)
And most importantly, they will lift your car up and carry it when your car (or other personal vehicle) is broken down. The average car weight range is 1.3-1.6 tons so they can lift and carry cars easily for long periods. 30 tons is just the capacity and they can lift up to 30 tons for emergency situations.
Selectable leverage/attachment points
Animal muscles are actually remarkably strong but can't contract very far, so they tend to operate at a mechanical disadvantage in order to give the required range of motion. We have fixed attachment points for our muscles on our skeleton, because that's what we evolved. The difference in strength between a chimp and a human isn't particularly the muscles, it's to do with the leverage those muscles can apply when attached to a skeleton.
With smart design, it could be possible to change the attachment points and hence the leverage for your artificial muscles relative to the joints of the skeleton. This could be as simple as winding out a lead-screw. Normally you'd never get to those ends of the range, but they're available.
Of course the trade-off is that you can't lift it very far! If your normal range of motion handles 100kg, then lifting a 30-tonne fire engine will give you 300 times less travel. A deadlift might normally move 60cm from a deep squat. Winding out your leverage to lift a fire engine scales that down to 2mm.
Still, you didn't specify how far it had to go, only that it had to be able to lift it. So 2mm still qualifies.
The robots are hard-wired to protect humans. They can't not do it. Being unable to protect a human drives a robot into an untenable mindset. They simply can't ignore a situation they're unable to assist with.
While out and about on an errand, a robot sees a car accident and there's a person pinned under an overturned vehicle. The robot analyses the situation, concludes that it doesn't have the servo-power to lift a car off a person and therefore cannot save this person.
Logically, they should either carry on with their business, or attempt to find a less direct way to help. But robots do not excel at second-order thinking. To a robot, the first-order problem is that the person is pinned under a vehicle, and that's the overriding most-important problem to solve. Everything else is just icing your cake before it's done.
So the robot is left with a dead-spot in its thinking. It can't simply ignore a human in distress, and it can't solve the problem.
So it stands there, stuttering, unable to think, unable to help, unable to support the people around it who need it to run and get a medkit...
Ultimately, a robot unable to help will have a mental breakdown as its priorities tear its artificial neurones apart. It'll have to be junked and a new artificial brain installed in the body.
Programming the deep-psychology of robots is extremely difficult and expensive, and the overriding priorities exist for good reasons. Nobody wants to mess with that. Nobody has yet succeeded in fixing this problem, not least because it's difficult to debug or reproduce in a lab.
So the solution is to give the robot the tools to solve 99.99% of physical problems it encounters.
Consequently, they have been designed with a very wide margin of physical capability to support this.
Under normal circumstances, they operate in human-norms. But in an emergency, if they need to be able to lift a car off someone and sprint three miles to the nearest hospital (carrying a 200lb adult human in their arms like a baby) they're perfectly capable of it.
Cost savings by using existing manufacturing facilities
Humanoid drones have been in use in many industries for several decades now, and the most common production models have high strength hydraulic musculature. When the AI breakthrough was made, the designers took an off the rack model to test out the consciousness installation so they wouldn't incur the extra costs of a bespoke design.
The robots are thick-skinned floating firefighters.
The robots are manufactured to be helpful in many scenarios - they have defibrillators (their logic airgapped from the robot itself, so offensive use is barred), fire extinguishers, ..., and are flotation devices and much more.
The flotation device usecase is coupled with the fire extinguisher role: the compressed gas used to propel the fire retardant can also be used to inflate the robot's arms (which has puncture resistant skin) to triple their size. To lift up a fallen object, the robot simply has to put itself, or an arm, under the object, and self-inflate, becoming a lifting bag.
Of course a murder-kill-death-robot could use this ability to smother people in enclosed spaces, or do any kind of damage to vital infrastructure, but realistically, any humanoid robot able to interact with human tools/artifacts can do that, absent an imperative to not do it. It is not an ability that endangers lives during a normal malfunction, though.