The Psychology of Robots Derek J. SMITH

The Psychologyof

Robots

Derek J. SMITHCentre for Psychology

University of Wales Institute, Cardiff

[email protected] http://www.smithsrisca.demon.co.uk

mailto:[email protected]?subject=Cymru-Arlein%20website%20design%20enquiry






http://www.smithsrisca.demon.co.uk/




An Educational Resource in UWIC's

"Psychology as a Career"Academic Video Series

DEDICATION

This presentation has been inspired by theWelsh Assembly Government's Reaching Higher - Reaching

Wider initiative for the modernisation of Welsh post-16 education, the H.E. Funding Council for Wales' FIRST

Campus initiative for promoting the university experience to groups who would not normally consider it an option,

and UWIC's own Widening Access Provision scheme.

As screened at the

2005 Christmas Lectureof the

South Wales Branch of the British Computer Society

Cardiff Bay Techniquest, 1st December 2005

PART 1

FROM AUTOMATA TO

ARTIFICIAL INTELLIGENCE

FROM AUTOMATA TO ARTIFICIAL INTELLIGENCE

This is probably humankind's first ever "robot".

It's a moving "tea-trolley" from around 3000 years ago. It would have helped "wait table" at feasts and festivals.

It was driven either by primitive clockwork or by a child-slave hidden in the base.


Here's an automaton from around 2000 years ago. The word comes from the Greek for "self activating".

Check out all the hidden pipes and linkages underneath!


This is an android, so called because it's human-shaped.

Mediaeval priests and alchemists tried to build androids to help them get rich, but they didn't know enough about technology or intelligence to succeed, so they often resorted to puppetry and clockwork instead.


This is the monk-alchemist Albert of Cologne. He spent 30 years working on a talking android.

Albert kept the workings of his 13th century android very secret, and nobody knows where it is now!

SOME PSYCHOLOGY .....

HERE IS ONE OF THE MOST IMPORTANT QUESTIONS EVER ASKED .....

IF YOU WERE ALBERT OF COLOGNE, AND YOU PUT ENOUGH GEARS AND HYDRAULIC

PIPES INTO A MECHANICAL PERSON, COULD YOU MAKE IT THINK?

FROM AUTOMATA TOARTIFICIAL INTELLIGENCE

So what is thinking?

The Spanish monk-philosopher Raymond Lull produced a "word processing" "computer" as long ago as 1274. Each disc turns independently, producing hundreds of combinations of brand new ideas.

Is this thinking, do you think?


Here's a typical pin-cylinder music box drive.

This one plays a tune - Ave Maria, say - on bells. See how the pins "know" the tune in advance and play it by each activating a corresponding hammer.

This is an early instance of "pre-programming". BUT IS IT THINKING?


Here's Blaise Pascal's calculating machine, from 1642.

When a right-hand wheel has advanced ten times, it advances its left-hand wheel once.

Our brains do this all the time when we have to "carry ten" in our sums. We call such things "thinking".

So is this machine thinking?


Here's another calculating machine, this time from the German philosopher-engineer Gottfried Leibniz in 1673. It's bigger and faster than Pascal's and driven by the handle front left. BUT IT'S STILL NOT THINKING!


In the late 18th century, the weaving industry found a way to automate their looms by scaling up earlier "music-box" technology. They brought pre-programming into industry!

This is the Jacquard punched card control system, on a 5 metre tall continuous loop, one per loom. BUT ONLY THE HUMAN OPERATOR IS DOING ANY THINKING.


Here's the bit that does our thinking for us.

To understand artificial intelligence, you need to know about the brain and how it works.

This area of study is known as "neuropsychology".


Fortunately, 18th century anatomists and physiologists were constantly discovering new facts about how the brain worked.

Here's a typical anatomy lesson.


And here's one of the most important physiological discoveries. This is Luigi Galvani. In 1786 he discovered that he could make dead frogs twitch by connecting them up to a battery!

This demonstrated the electrical nature of the nerve impulse and got everybody wondering whether Descartes had been right about animals being just fancy machines.


Around this time, calculating machines were getting a lot bigger.

In 1822 the British inventor Charles Babbage designed a giant calculating machine called a "difference engine". Unfortunately, it took a long time to build. This is how far he'd got by 1832 .....


Between 1832 and his death in 1871, Babbage worked on a much more powerful machine called an "analytical engine".

It was intended to read in pre-programmed instructions from Jacquard cards.

Unfortunately, Babbage never finished this machine due to shortage of funds.


This is the Jacquard punched card instruction system Babbage wanted to use to program his all-purpose computer.


..... and this is Ada, Lady Lovelace.

She helped Babbage "program" his new machines, and is therefore widely acclaimed as the World's first "computer programmer"


In 1834, the French physicist André Ampère introduced the term cybernétique to describe all forms of automatic control mechanisms.

We know the word today as "cybernetics", and have used it for fictional beings such as cybermen, cyborgs, etc.

Daddy!

That's my boy!


So let's summarise how far we'd got in the mid-1800s.

Thanks to the anatomists, we knew something about the brain.

Thanks to clockmakers and weavers we had programming.

Thanks to Babbage and Lovelace we had computing.

Thanks to Ampère we had cybernetics.


In 1865, the French novelist Jules Verne published "From the Earth to the Moon". Here's the inside of his imaginary space capsule. It doesn't seem to contain a calculator.

As we'll be seeing later, NASA now proposes to crew most of their future spacecraft with artificial intelligence systems and androids, and not take any humans.


By 1893, androids, automata, and data processing were pretty much everyday topics of conversation.

This android is called Boilerplate. He liked posing for photos with celebrities, but there wasn't a lot going on inside his head.

However, there was also a darker side to automata - as weapons of war .....


Intelligent weapons were experimented with during World War 1. This is the US Army's "Kettering bug" aerial torpedo, from 1918. In fact the only artificial intelligence it had on board was a gyroscope system for fine trim. The rest was provided by a human operator by radio control.


The word "robot" came from the 1923 play "R.U.R." by Karel Capek. He described a world run by imaginary artificial workers known as "robotniks".

This robotnik, in case you hadn't noticed, is just an actor in a silver suit!

What do you mean machines don't need

teabreaks!


In 1932, the Welsh physicist Charles Wynn-Williams invented an electronic adding machine. It worked just like Babbage's calculating engine, only many times faster because it had no moving parts. It was the prototype of the sort of logic circuits used in modern digital computer.


Remember this?

If Galvani could make frog nerves behave like wires .....


..... what if we connected up some frog nerves according to Wynn-Williams' circuit design?

Would frog legs be able to add up, perhaps?

Surprisingly, the answer is probably yes!!


When the Second World War started in 1939, military researchers soon brushed the dust off the idea of intelligent weapons. This is the Luftwaffe's Henschel 293 "stand-off" bomb, from the early 1940s. It sank a lot of ships and killed a lot of sailors.


In 1941, the science fiction writer Isaac Asimov invented the "laws of robotics", as follows:

(1) A robot may not injure a human being, or, through inaction, allow a human being to come to harm. (2) A robot must obey, unless this conflicts with (1). (3) A robot must protect its own existence, unless this conflicts with (1) or (2).

So it's not really fair to describe robot weapons as robots, after all, because they don't have any free will.

MORE PSYCHOLOGY .....

BUT WHAT IS FREE WILL ANYWAY?

PSYCHOLOGISTS DON'T REALLY KNOW WHAT FREE WILL IS IN HUMANS, SO HOW

ARE WE GOING TO DESIGN IT INTO INTELLIGENT MACHINES?

PART 2

MODERN ARTIFICIAL INTELLIGENCE


In 1943, British wartime scientists invented the first programmable electronic computer to help them decode the enemy's military signals.

Much of the construction work was done by a team of telecommunications engineers led by Tommy Flowers (shown here in the 1990s).


This is a replica of the machine Flowers and his team succeeded in putting together.

It's name was Colossus, and (like Albert of Cologne's android) its design was kept colossally secret.


After the war, computers began to re-shape the world. Here's a British Tabulating Machinery HEC computer, vintage 1953.


Cinema and TV loved robots to pieces. Here are a couple of the nicer inventions. These were very intelligent, but only because they had actors inside.

Hi, I'm Robby, and this is my friend Twiki.


..... and here's one of the nastier ones (although he turned out to be a bit of a softie on the quiet).

I swear I will not hurt you. (But if you're lying, I'll be

back!)

MODERN ARTIFICIAL INTELLIGENCE(BASIC RESEARCH)

So it's important to realise that looks can be deceiving. This is Johnny Five, from the 1986 movie "Short Circuit". He looks bright enough, but he was worked almost entirely by remote control. Just clever puppetry yet again!


This, on the other hand, is Kismet, one of a range of state-of-the-art research projects being carried out at MIT. Kismet can mimic a number of human emotional expressions, and is especially good at those involving eyebrows!


This is the Honda-Kawasaki ASIMO. It can walk, has colour vision, face recognition, gesture recognition, and speech, and can even obey simple commands. But it has little intelligence, and no free will at all!

Hi, I'm ASIMO. I was born in 2003, and I can walk (but only when they tell me to).


QRIO is Sony's answer to ASIMO. It can do most things ASIMO can do, but has better hand and finger control.

Here we see four identical QRIOs playing Ave Maria on handbells.

So are they just fancy musical boxes, or metal people?

MODERN ARTIFICIAL INTELLIGENCE(AT WAR)

Here's a fierce looking robot. This one takes its own life at the end of its mission! If it was "alive" this would arguably make it a metal suicide bomber!

I shan't be home for tea, tonight, darling!

MODERN ARTIFICIAL INTELLIGENCE(IN THE OPERATING THEATRE)

Here's a robot neurosurgeon performing a little brain surgery.

MODERN ARTIFICIAL INTELLIGENCE(IN THE FACTORY)

Robots can be a valuable part of the process of automation. Here's a robot on a factory assembly line.

It takes a lot of psychology to get the machines and the people to work efficiently alongside each other

Hi, I'm Roboblaster from Rossler, Inc.

MODERN ARTIFICIAL INTELLIGENCE(READING THE NEWS)

Here's a robot made only out of spots of light on a display screen. She's a "software agent" - a virtual android with lots of intelligence and no complicated mechanical bits at all.

She's the BBC's Ananova, prototype "cyberbabe" and virtual newsreader.

And now over to Westminster .....

MODERN ARTIFICIAL INTELLIGENCE(AT THE RECEPTION DESK)

This is Valerie, the android robot receptionist under trial at Carnegie Mellon University in the US.

She's going to be as near lifelike as they can make her! She'll even remember what you talked about last time you visited!

So how's your Bert's lumbago?

MODERN ARTIFICIAL INTELLIGENCE(IN THE GARDEN)

Here's a robot from Friendly Robotics Inc. for mowing the lawn.

MODERN ARTIFICIAL INTELLIGENCE(IN THE SITTING ROOM)

And here's Electrolux's latest robot vacuum cleaner, the Trilobite.

PART 3

WHERE TO NEXT?

WHERE TO NEXT?(ARTIFICIAL COMPANIONS)

Remember that to be an "android" you have to have the shape of a man (or woman).

Hi, I'm Valerie, the most advanced android in the

world, from Android World, Inc.

But I'm not such a babe without my

face on ....

WHERE TO NEXT?(SPACE EXPLORATION)

Not all robots will be shaped like humans (so they'll be robots, but not androids).

This prototype NASA rover has six legs because it needs to cope with rough ground on alien planets.


However, because humans and robots might need to share a space capsule one day, NASA will probably make android robots as well.

This one's called Robonaut. He won't need oxygen or food, but it makes sense for him to share tools with the humans, so he needs lifelike arms and hands.

Pass me that hammer, Dai.


Android robots will be easier to identify with - form friendships, even. Or perhaps one day fall in love!!!

Psychologists will have to investigate the problems of long-term robot-human confinement before any deep space mission can take place - which do you think will go mad first?

Robots have feelings too, you know! (In fact, I've got 150 sensors in each arm.)


Psychologists are also working on Robonaut's brain, because intellectually he's still a long way from being ready for the trip.

(To put it bluntly, he's still a bit dull!)

"Two plus two is - er - five?

MORE PSYCHOLOGY .....

HOW WILL ROBOTS AND ANDROIDS AFFECT HOW WE LIVE AND WORK?

WILL THEY EVER "THINK" FOR THEMSELVES?

WILL THEY GET BORED?

WILL THEY NEED TO GO TO CHURCH?

ETC., ETC., ETC.

NOBODY KNOWS FOR SURE

BUT BECOME A PSYCHOLOGIST AND BE ONE OF THE FIRST TO FIND OUT!!

THE END

(OR IS IT?)

FOR FURTHER INFORMATION

For leaflets on careers in psychology, to book your place on an Open Day, or just for general advice on the university experience as a whole, please contact .....

Mrs Helen Campling, Centre for Psychology, UWIC, Western Avenue, Cardiff, CF5 2YB

Tel: 02920 417210 E-mail: [email protected]

Thank you for watching"The Psychology of Robots"

This has been an educational resource in UWIC's

"Psychology as a Career"Academic Video Series

© 2005, University of Wales Institute, Cardiff

The Psychology of Robots Derek J. SMITH

Documents

modern artificial

automata toartificial

artificial

space exploration

calculating

basic research

wales institute

educational