- What if robots were able to interact with humans in every aspect of our daily lives?

Robots could work in our hospitals, in our homes, in our workplaces.

They could do manual labor or go on challenging search and rescue missions.

And what if these robots didn't just work with us, but acted and looked like us?

Would we be able to integrate them into society in a more meaningful and impactful way?

And what's stopping us from doing so?

I'm Sinead Bovell.

Let's talk robot design.

[cheerful music] This is one of the first humanoid robots in science fiction.

His name is Tik-Tok.

No, not that TikTok.

He's a friend and companion of Dorothy in "The Wizard of Oz" and he kind of looks like a steampunk version of Mario.

In the 120 years or so between in Tik-Tok and, well, TikTok, scientists and engineers have been obsessed with creating robots that are built just like us.

For years, roboticists have been building all sorts of machines that can handle jobs humans might not want to do.

The Ds, dull, dirty, dangerous, difficult.

The estimated number of jobs that robots could replace varies, but most economists agree it's an accelerating trend, with some thinking that in the future, upwards of 78% of manufacturing jobs could be done by robots.

But why are some roboticists building humanoid robots in particular?

Why do they have to be built like us?

It turns out there are some very practical reasons.

For starters, robots will be entering into a world that we've optimized for us, humans.

Stairs, tools, home appliances, planes, buses, cars, all designed for us, which is why for now, it's probably easier to design robots that are built like us instead of redesigning our world for robots.

- Why do we need humanoid robots?

Well, I have a dream.

In the future, I would love to live with robots in our house, robots doing the dishes, taking out the trash.

Unless the robot is a humanoid robot, it won't be able to navigate this environment or even use tools designed for humans.

- This is Dennis Hong, a mechanical engineer and roboticist working on creating a future filled with robots, and he's made a lot of them.

- The famous architect Louis Sullivan once said form follows function, which means that the shape of an object is dictated by what it really needs to do.

You can think of humanoid robots as a Swiss Army Knife.

It's not a single tasker, not a specific robot that only does one thing, but if it's a humanoid robot, it's a multitasker.

It can do basically what a human can do.

- Scientists and engineers steal from nature all the time.

In the 1940s, George de Mestral, a Swiss engineer, spent hours plucking burrs out of his dog's fur after long hikes.

He became fascinated with nature's ability to develop the perfect hook that could latch onto cloth, fur, and more.

A few years later, Velcro was born.

And that's just one of many examples.

Trains have been designed after birds' beaks, wind turbines from the flippers of whales, solar cells from a butterfly's wings.

Scientists are even exploring how we could build future cities to be more like termite mounds with naturally evolved cooling and heating design features that would reduce energy costs by up to 90%.

It's a concept called biomimicry, and it makes a lot of sense.

After a few billion years of evolution, nature has gotten a few things right.

So are there parts of the human body that we should mimic in robots?

- So the human body is amazing.

We roboticists try to mimic, try to even get close to what humans can do, but we've still got a long way to go.

Now of course for the brain, you can say that we have a computer.

For eyes, we have a camera.

But what about the muscles, right?

Most robots use electric motors with gears, but the thing is, these electric motors with gears which are called servomotors, they move, but they move very differently than biological muscles.

So still, this is one of the holy grails in robotics, and a lot of researchers are working on that.

- But what are the downsides of humanoid robots?

Will they create jobs or eliminate them?

Is there actually cause for concern with creating robots that are built too much like us?

What are the things that could go wrong?

These are questions that researcher and ethicist Kate Darling grapples with on a daily basis.

- It's hard to predict the future.

I think right now, people both overestimate and underestimate what the technology can do.

My biggest concern would be that this is happening within an environment of unbridled corporate capitalism.

I think the concern is not so much the robots themselves as what our political and economic systems are going to incentivize in terms of how we use them.

- And then there's Sophia.

Sophia's what's called a social robot, and it might not look like it, but this is a pretty different approach to designing humanoid robots.

The bots being made by Dennis Hong and others are machines meant to interact with objects, other robots, complete tasks, and their design reflects that.

But Sophia is meant to interact with people.

She's what's called an embodied AI chatbot, and she could be used for a wide swath of purposes.

Some roboticists are building social robots to assist with elderly care, which could handle tasks like medication reminders, monitoring vital signs, detecting falls, and more.

These care bots could make nursing homes more efficient and more affordable.

Social robots could also help teach young students in the form of classroom tutors.

They could function as personal assistants, or act as emotional support tools for people who just need somebody to talk to.

But do social robots have to look like humans?

- I do think sometimes we fall into the trap of thinking that we have to develop robots in a humanoid form for everything.

There's a robot, it is a little bit controversial.

The PARO baby seal robot.

It's a medical device.

It's used in nursing homes with dementia patients.

It really gives people the sense of nurturing something, which yeah, that's kind of creepy.

We're giving people this sense of nurturing something that's not alive.

But on the other hand, it's a really important experience for people to have whose lives have been kind of reduced to being taken care of by others.

- Social robots don't always have to be designed like humans to be effective, and some research even shows that a humanoid robot can lead to a more disappointing experience for people because they often have expectations about how a robot should behave if it looks like a human.

But if we do want to design social robots, or any robot, to resemble humans, they have to get past the uncanny valley.

You've probably heard of the uncanny valley.

As a robot becomes more lifelike and realistic, we tend to respond better to them.

That is, until they get too close to being real, not quite lifelike, but just right there.

That's when they really freak us out.

But why do our brains do this?

The original concept was introduced by a Japanese robotics professor, Masahiro Mori, and it wasn't a research paper in a scientific journal, it was a hypothesis in an essay he wrote in 1970 appropriately titled "The Uncanny Valley".

But that wasn't enough for researchers who wanted to understand why our brains work that way.

So they tried to come up with answers, which spread a wide gamut.

Some researchers think it reminds our subconscious of a dead body, and causes anxiety, a concept called mortality salience.

The idea is that when we're faced with the inevitable fact that we're going to die, it naturally freaks us out.

But other researchers think that our brains are literally telling us two different things.

This is a robot!

No, this is a human!

And that cognitive dissonance overloads our brain, and we shut down and react negatively.

But this isn't a universal problem.

Different cultures, different demographics, different age groups all react differently to the uncanny valley, and some researchers dismiss the idea entirely.

Robots offer some pretty exciting and compelling advantages for humans, but as with every technology, there are always important ethical considerations we need to think about.

For standard humanoid robots, designing them exactly like humans could leave us vulnerable to competing with robots in the workforce instead of using them to supplement human skills.

And for social robots, while they can offer new and beneficial relationships for humans where human care may not be accessible, there could also be lots of opportunities for manipulation by the companies that are building them.

Imagine a world of microtransactions, paying for upgrades and repairs, unlocking new features, all for a robot that you've developed a real emotional attachment with.

The potential for exploitation is high.

I think the more we think of robots as partners for us in what we're trying to achieve instead of replacements for us, the better off we'll be in designing and building robots for the benefit of society.