Full interview: "Godfather of artificial intelligence" talks impact and potential of AI

Summary

このビデオのスクリプトでは、AIと機械学習の現在の重要な時点について語られています。特に、ChatGPTのような大規模言語モデルの可能性と、一般公衆の反応の大きさに注目しています。ニューラルネットワークと後方伝播の概念、ディープラーニングの進化、そしてこれらの技術が人間の脳の理解にどのように貢献するかに焦点を当て、AI技術の将来への展望を提供しています。また、これらの進歩が社会、職場、倫理的問題にどのように影響を与えるかについても考察しています。

Takeaways

• 🔍 AIのこの瞬間は、言語モデルが素晴らしいことができるという事実を一般の人々が突然認識した画期的な瞬間です。
• 🤖 チャットGPTに対する一般の反応は、研究者たちも驚くほど大きなものでした。
• 🧠 AI分野には、論理と推論に基づく主流のAIと、生物学的なニューラルネットに基づく別の学派が存在しています。
• ⚙️ ニューラルネットワークが1980年代にうまく機能しなかったのは、コンピュータの速度とデータセットの大きさが不十分だったからです。
• 💡 ニューラルネットワークによる学習は、脳の働きを模倣することに基づいており、それは人間が読み書きや数学を学ぶ能力にも関連しています。
• 📈 ディープラーニングの進化は、複数の表現層を持つニューラルネットを用いて複雑な学習を可能にする重要な転換点でした。
• 🖥️ バックプロパゲーションとは、予測が間違っていた場合にネットワークの各接続の強度を調整して、より正確な予測を行うプロセスです。
• 🔬 画像認識システムの進化は、物体を正確に認識できるようになったことで、AIの重要な成果の一つとなりました。
• 🤔 AIが言語を理解するプロセスは、単に次の単語を予測すること以上のものであり、実際の理解を必要とします。
• 🚀 AI技術の発展は、産業革命や電気の発明と同じくらいの規模で人類の生活に影響を与える可能性があります。

Q & A

ChatGPTに対する一般の人々の反応について、最初に使用したときの反応はどうでしたか？

ChatGPTそのものはそれほど驚きませんでしたが、一般の人々の反応の大きさには多少驚きました。

AIとニューラルネットワークにおける2つの主要な学派とは何ですか？

AIには「メインストリームAI」があり、論理と推論に基づいていました。もう一方の学派は「ニューラルネットワーク」で、生物学的な学習方法、すなわちニューロン間の接続の変化に重点を置いていました。

ニューラルネットワークが1980年代にうまく機能しなかった主な理由は何ですか？

コンピューターの処理能力が十分ではなく、データセットも十分に大きくなかったためです。

ChatGPTの訓練に用いられる「バックプロパゲーション」について、脳が同じ方法を使用しているとは思いますか？

いいえ、脳がバックプロパゲーションを使用しているとは思わず、AIと脳の学習方法には根本的な違いがあると考えています。

2006年にAI研究で起きた重要な変化とは何ですか？

深層学習（ディープラーニング）の研究が始まり、複数のレイヤーを持つニューラルネットワークが複雑な学習を行うことができるようになりました。

画像認識システムにおいて、ニューラルネットワークのアプローチが従来の方法とどのように異なるか？

ニューラルネットワークはランダムな重みから始めて、エラーを最小限に抑えるように重みを調整することにより、自動的に特徴検出器を学習します。従来の方法では、人が手動で特徴を定義していました。

ニューラルネットワークの研究が加速した理由は何ですか？

コンピュータの処理能力の向上とデータの量が増加したため、ニューラルネットワークをより大きく、より複雑にすることができるようになりました。

ニューラルネットワークと人間の脳との間の主要な違いは何ですか？

ニューラルネットワークはデジタルコンピュータ上で動作し、高い計算能力を必要としますが、人間の脳はアナログで、はるかに低いエネルギーで動作します。

将来、コーディングのスキルが必要かどうかについての見解は？

AIの発展により、コーディングの仕事の性質が変化する可能性がありますが、創造的な側面により重点を置くようになるかもしれません。

AIと脳の研究が異なる道を進むと考える理由は何ですか？

AIの発展はバックプロパゲーションに依存していますが、人間の脳が同じメカニズムを使用しているとは思わないため、理論的に異なる方向に進んでいると考えています。

Outlines

00:00

🤖 AIの発展とチャットGPTの影響

現在のAIの発展と機械学習の状況について議論し、チャットGPTがどのように驚くべき成果を上げ、大衆に注目されるようになったかについて説明しています。また、Microsoftがリリースしたことにより、チャットGPTの存在が広く知られるようになった点も触れています。

05:01

🧠 AIと脳の学習の比較

AIと脳の学習プロセスの比較、特にニューラルネットワークと脳の機能の類似性について説明しています。また、過去のAIの理論と現在までの発展、そしてその途中で起こった重要な出来事についても触れられています。

10:04

🖼️ 画像認識技術の進歩

画像認識技術の発展と、特に深層学習がその分野に与えた影響について説明しています。また、大容量データベースの構築と、それらを用いた学習プロセスでの成果についても触れられています。

15:06

🌐 AI技術の将来展望

AI技術の将来的な展望、特に低消費電力での運用や、AIと人間の関係性について議論しています。また、AIがもたらす政治的・経済的課題についても言及されています。

20:07

🚀 AIの発展速度とその影響

AIの発展速度が加速していることと、それが社会に与える可能性について述べています。また、AIがもたらす脅威や、それを管理するために必要な取り組みについても話し及んでいます。

25:07

🛠️ AI研究におけるカナダの役割

カナダにおけるAI研究の発展と、どのようにしてその分野のリーダーシップを確保したかについて説明しています。また、研究資金の提供方法や、研究者の交流を支援する取り組みについても触れられています。

Keywords

💡チャットGPT

チャットGPTは、大規模な言語モデルの一つであり、このビデオの主要テーマです。一般大衆にAIの進歩を広く認識させるきっかけとなった技術です。スクリプト内で、チャットGPTが公開された際の大きな反響について触れられており、それがどのようにしてAI技術の「ピボタルモーメント」を形成したかが語られています。

💡ニューラルネットワーク

ニューラルネットワークは、脳のニューロンのネットワークを模倣したコンピュータアルゴリズムです。このビデオでは、AIと機械学習の発展におけるニューラルネットワークの重要性が強調されています。特に、過去のAI研究がニューラルネットワークに焦点を当てる前の対立するアプローチや、ニューラルネットワークが現在どのようにして学習と認識を可能にしているかについて述べられています。

💡ディープラーニング

ディープラーニングは、多層ニューラルネットワークを使用して複雑なパターンを学習する技術です。このビデオでは、2006年頃から始まったディープラーニングの時代がAI研究における転換点であったことが語られています。特に、音声認識や画像認識におけるその成功が触れられ、これがどのようにして技術の発展に寄与したかが説明されています。

💡バックプロパゲーション

バックプロパゲーションは、ニューラルネットワークの学習において誤差を逆伝播させることで重みを調整する方法です。スクリプトでは、このアルゴリズムがどのようにして初期のニューラルネットワークの問題を克服し、学習過程を可能にしたかが説明されています。特に、画像認識タスクにおけるその応用例が詳細に語られています。

💡トランスフォーマー

トランスフォーマーは、自己注意メカニズムを用いたモデルで、特に言語処理タスクに革命をもたらしました。ビデオでは、トランスフォーマーがチャットGPTの基盤技術であり、その発展がどのようにしてディープラーニング研究における新たな時代を開いたかが語られています。

💡人工知能の倫理

人工知能の倫理は、AI技術の発展と応用に伴う倫理的、社会的な問題を扱います。ビデオでは、AIの影響について深く考察する部分があり、特に真実性、自律武器、仕事の未来など、AI技術が人間社会に与える潜在的な影響について議論されています。

💡ジェネラルAI

ジェネラルAI（汎用人工知能）は、あらゆる知的タスクを人間と同等以上にこなせるAIのことを指します。ビデオでは、このようなAIの実現が近づいているかもしれないという懸念が表明され、その潜在的な危険性についても言及されています。

💡データセット

データセットは、機械学習モデルを訓練するために使用されるデータの集合です。ビデオの中で、過去には限られたデータセットでの学習が挑戦であったことや、現代の大規模データセットがAIモデルの性能向上に貢献していることが説明されています。

💡計算能力

計算能力は、AIモデルを訓練し実行するために必要なコンピュータの処理速度やメモリ容量を指します。このビデオでは、過去のAI研究が限られた計算能力によって制約されていたが、現代の高性能コンピューティングによってこれらの制約が克服されたことが語られています。

Highlights

ChatGPT's impact demonstrates the capabilities of large language models, surprising many with its rapid adoption.

Early AI models like GPT-2 and Google's joke-explaining model showcased the potential of AI in natural language understanding.

The public's strong reaction to ChatGPT highlighted the growing interest and potential applications of AI technology.

AI's evolution from focusing on logic and reasoning to neural networks and learning through connections.

The significance of computational power and data availability in the advancement of neural networks since the 1980s.

The role of deep learning in enabling AI to tackle complex tasks like speech recognition and image identification.

Backpropagation's role in improving AI's learning capabilities, despite doubts about its similarity to brain functions.

The transformation in object recognition through neural networks, leading to significant improvements over traditional AI methods.

The emergence of generative models and transformers as key technologies in advancing AI capabilities.

The ongoing challenge of creating AI that can understand and replicate human reasoning and creativity.

The potential for AI to transform various industries and occupations, complementing human creativity and efficiency.

The necessity of considering ethical implications and governance in the development and deployment of AI technologies.

The debate over AI's understanding of truth and its ability to reconcile different worldviews.

The potential risks and benefits of AI in autonomous weapons and the importance of global cooperation in governance.

The foundational role of curiosity-driven research in AI's development and the significance of basic research funding.

The philosophical and practical questions surrounding AI sentience and its implications for society and technology.

Transcripts

how would you describe this current

moment in AI machine learning whatever

we want to call it

I think it's a pivotal moment chat GPT

has shown that these big language models

can do amazing things and the general

public has suddenly caught on yeah we

have because Microsoft released

something

and they're suddenly aware of stuff that

people of the big companies have been

aware of for the last five years yeah

what did you think the first time you

used chat GPT

um

it's well I've used lots of things that

came before chat gpg that were quite

similar so chat GPD itself

didn't amaze me much gpt2 which was one

of the earlier language Brands amazed me

and a model at Google amazed me that

could actually explain why a joke was

funny oh really yeah in just natural

language it'll tell you yeah you tell it

a joke not for all jokes but for quite a

few of them it can tell you why it's

funny okay and that you it seems very

hard to say it doesn't understand when

it can tell you why a joke's funny so if

chat GPT wasn't

all that surprising or impressive were

you surprised by The public's reaction

to it because the reaction was big

yes I think everybody was a bit

surprised by how big the reaction was

that it was the sort of fastest growing

up ever yeah

um

maybe we shouldn't have been surprised

but people the researchers had kind of

got used to the fact that these things

actually worked yeah you were famously

like

half a century ahead of the curve on

this AI stuff go ahead correct me go

ahead not really not really because

there were two schools of thought in AI

um there was mainstream Ai and then

there was years but it was all about

friends yeah that thought it was all

about reasoning and logic and then there

was neural Nets which won't call AI then

um

which thought that you better study

biology because those were the only

things that really worked

and so mainstream AI based its theories

on reasoning and logic and we based our

theories on the idea that connections

between neurons change and that's how

you learn

and it turned out in the long run

um

we came up trumps

um but in the short term it looked kind

of hopeless well looking back knowing

what you know now do you think there's

anything you could have said then that

would have convinced people

I could have said it then but it

wouldn't have convinced people and what

I could have said then is the only

reason that neural networks weren't

working really well in the 1980s was

because the computers weren't fast

enough and the data sets weren't big

enough

but back in the 80s

the big issue was could you expect a big

neural network with lots of neurons in

it compute nodes and connections between

them that learns by just changing the

strengths of the connections could you

expect that to just look at data and

with no kind of innate prior knowledge

learn how to do things and people in

mainstream AI thought that was

completely ridiculous it sounds a little

ridiculous it is a little ridiculous but

it works

and how did you know or why did you

Intuit that it would work because the

brain works because that's you have to

explain how come we can do things and

how come we can do things like we didn't

evolve for like reading reading's much

too recent for us to have had

significant evolutionary input to it but

we can learn to do that and Mathematics

we can learn that so there must be a way

to learn in these neural networks

yesterday Nick Frost who used to work

with you told us that you are not really

that interested in creating AI your core

interest is just in understanding how

the brain works yes I'd really like to

understand how the brain works obviously

if your failed theories of how the brain

works lead to good technology you cash

in on that and it gets get grants and

things but

um I really would like to know how the

brain works and I think there's

currently a Divergence between the

artificial neural networks that are the

basis of all this new Ai and how the

brain actually works I think they're

going different routes now so we're

still not

going about it the right way

that's what I believe this is my

personal opinion but all of the big

models now use a technique called back

propagation which you helped popularize

popularize in the 80s very good

um and I don't think that's what the

brain is doing explain why okay there's

a fundamental difference between two

different there's two different paths to

intelligence

so One path is a biological path where

you have Hardware that's a bit flaky an

analog so what we have to do is

communicate by using natural language

also by showing people how to do these

imitation and things like that but

instead of being able to communicate a

hundred trillion numbers we can only

communicate what you could say in a

sentence which is not that many bits per

second yeah and so we're really bad at

communicating compared with these

current computer models that run on

digital computers it's almost infinite

they're able to that's a communication

band with this huge yeah because they're

exactly the same model they're clones of

the same model running on different

computers and because of that they can

see huge amounts of data because

different computers can see different

datas and then they can combine what

they learned more than any person could

ever comprehend far more than any person

could have become and yet somehow we're

smarter than them still okay so they're

like idiot savants right

chat GPT knows much more than any one

person if you had a competition about

you know how much you know it would just

wipe out any one person it was amazing

at bar trivia yes it would do amazing it

would do me and it can do all you can

write poems it can

you know

um they're not so good at reasoning

we're better at reasoning we have to

extract our knowledge from much less

data so we've got a hundred trillion

connections most of which we learn but

we only live for a billion seconds which

isn't very long

whereas things like chat GPT

have run for much more time than that to

absorb all this data but on many

different computers

1986 you publish a thing in nature that

is the idea we're going to have a

sentence of words and it'll predict the

last word yes that was the first

language model that's basically what

we're doing now yes and no 1986 was a

long time ago why still did people not

say oh okay I think he's on to something

oh because back then if you asked how

much data I trained that model on I had

a little

um simple world of just family

relationships there were 112 possible

sentences and I trained it on 104 of

them and checked out whether it got the

last eight right okay and how would it

do

it got most of the last eight right okay

it did better than symbolic AI so it's

just that the computers weren't powerful

enough at the time the computers we have

now are millions of times faster they're

parallel but they can do millions of

times more competition so I did a little

computation if I'd taken the

um

computer I had back in 1986

and I started learning something on it

it would still be running now and not

have got there huh

um and that's stuff that would now take

a few seconds to learn did you know

that's what was holding you back

I didn't know it I believe that might be

what was holding us back

but people sort of made fun of the idea

that

the claim that well you know if I just

had a much bigger computer and much more

data everything would work and the

reason it doesn't work now is because we

haven't got enough data on enough

compute that's seen as a sort of lame

excuse for the fact that your thing

doesn't work

was it hard in the 90s doing this work

in the 90s computers were improving but

um yes so there were other learning

techniques that are on small data sets

worked at least as well as neural

networks and were easier to explain and

had much fancier mathematical theory

behind them

and so people within computer science

lost interest in neural networks within

psychology they didn't because within

psychology they're interested in

how people might actually learn and

these other techniques looked even less

plausible than back propagation here

which is an interesting part of your

background you came to this not because

you were interested in computers

necessarily but because you were

interested in the brain yes I sort of

decided I was interested in Psychology

originally then I decided we were never

going to understand how people work

without understanding the brain the idea

that you could do it without worrying

about the brain that was a sort of

fashionable idea back in the 70s but

um I decided that wasn't on you had to

understand how the brain worked so we

fast forward now to the 2000s is there a

key moment you think back to is a

turning point when it's like okay

our side is going to Prevail in this

around 2006

we started doing what we call Deep

learning

um before then it had been hard to get

to neural Nets with many layers of

representation

to learn complicated things

and we find better ways of doing it

better ways of initializing the networks

called pre-training and the p in chat

gbt stands for pre-training okay

um and the t is Transformer and G is

generative and it was actually

generative models provided this better

way of pre-training neural ads so the

seeds of it were there in 2006 in by

2009 we'd already produced something

that

was better than the best speech

recognizers and recognizing which

phoneme you were saying using different

technology than all the other speech

recognizers were then then the standard

approach which you've been tuned for 20

for 30 years there were other people

using neural Nets but they weren't using

deep neural ads and then there's a big

thing happens in 2012. yes to actually

two big things okay one is that the

research we'd done in 2009 done by two

of my students over a summer that led to

better speech recognition

that got disseminated to all the big

speech recognition Labs that Microsoft

and IBM and Google

and in 2012 Google was the first to get

into a product and suddenly speech

recognition on the Android became as

good as Siri if not better

so that was a deployment of deep neural

Nets applied to speech recognition three

years earlier at the same time as that

happened within a few months of that

happening

two other students of mine

developed an object recognition system

that would look at images and tell you

what the object was and it worked much

better than previous systems how did

this system work okay there was someone

called Faith a Lee in her collaborators

that created a big database of images

like a million images of a thousand

different categories you'd have to look

at an image and give your best guess

about what the primary object was in the

image so the images would typically have

one object in the middle yeah and I

didn't have to say things like bullet

train or Husky or and

the other systems

were getting like 25 errors and we were

getting like 15 errors okay within a few

years that 15 went down to three percent

which was about human level and can you

explain in a way people would understand

the difference between the way they were

doing it and the way your team did it

I can try that's all we can hope for

okay

so suppose you wanted to recognize a

bird in an image okay the image itself

let's suppose it's a

200 by 200 image

that's got 200 times 200 pixels

and each pixel has three values for the

three colors RGB

and so you've got 200 by 200 by 3

numbers in the computer it's just

numbers in the computer right

and the job is

to take those numbers in the computer

and convert them to a string that says

bird

so how would you go about doing that and

for 50 years people in standard AI tried

to do that and couldn't

got a bunch of numbers into a label that

says bird

so here's a way you might go about it

at the first level of features you might

make feature detectors things that you

take little combinations of pixels okay

so you might make a feature detector

that said look if all these pixels are

dark and all these pixels are bright I'm

going to turn on okay and so that

feature detector would represent an edge

here okay a vertical Edge you might have

another one that said if all these

pixels are bright and all these picks as

a dark I'll turn on that would be if

each detector that represent in a

horizontal Edge okay and you can have

others for edges of different organs we

had a lot of work to do all we've done

is made a box right so we've got to have

a whole lot of feature settings like

that and that's what you actually have

in your brain okay so if you look in a

cattle monkey cortex it's got feature

detectors like that

um then at the next level

you might say if you were worried up by

hand you would create all these little

feature detectors at the next level you

would say

um okay suppose I have two two Edge

detectors that join at a fine angle

that could just be a beak so the next

level up will have a feature detector

that detects two of the lower level

detectors joining a fine angle okay

we might also

notice a bunch of edges that sort of

form a circle we might have a detector

for that okay then the next level up we

might have a detector that says hey I

found this beak like thing and I find a

circular thing in roughly the right

spatial relationship to make the eye and

the beak of a bird

and so at the next level up you'd have a

bird detector that says if I see those

two there I think it might be a bird

okay and you could imagine wiring all

that up by hand okay and so the idea of

back propagation is

just put in random weights to begin with

and now the featured textures would just

be rubbish whether it be garbage okay

okay but look to see what it predicts

and if it happened to predict bird it

wouldn't but if it happened to leave the

weights alone

um you got it right the connection

strings but if it predicts cat

then what you do is you go backwards

through the network and you ask the

following question

and you can ask this with a branch of

mathematics called calculus but you just

need to think about the question and the

question is how should I change this

connection strength so it's less likely

to say cat I'm more likely to say bird

that's called the ER the error the

discrepancy right okay and you figure

out for every connection strength how I

should change a little bit to make it

more likely to say bird and less likely

to say cat and a person's figuring that

out or the algorithm is set to work a

person has said this is a bird so a

person looked at the image and said it's

a bird it's not a cat it's a bird so

that's a label supplied by a person

but then the algorithm back propagation

is just a way of figuring out how to

change every connection strength to make

it more likely to say burden less likely

to say cat it just keeps trying keep

turning it just keeps doing that and now

if you showed enough birds and enough

cats when you showed a bird it'll say

burden when you showed a cat it'll say

cat and it turns out that works much

much better than trying to wire

everything by hand and that's what your

students did on this image database

that's why they did on the image check

device yes and they got it to work

really well now they were very clever

students in fact one of them Ilya

sutskova

is also one of the main people buying

chat gbt

so that was a huge moment in Ai and chat

gbt was another huge moment and he was

actually involved in both of them yeah

yeah

I don't know maybe it's cold in the room

you got to the end of the story I go

shivers

the idea that you do this little dial

thing and it says bird it feels like

just an amazing breakthrough yeah I it

was

um mainly because

the other people in computer vision

thought okay so these neural Nets they

work for simple things like recognizing

a handwritten digit but that's not a

real complicated image with sort of

natural background with stuff it's never

going to work for these big complicated

images and then suddenly it did

and to their credit the people who've

been really staunch critics of neural

Nets and said these things are never

going to work when they worked they did

something that scientists don't normally

do which she said oh it worked we'll do

that people see it as a huge shift yes

it was quite impressive that they

flipped very fast because they saw that

it worked better than what they were

doing yeah you make this point

that when people are thinking both about

their machines and about ourselves in

the way we think we think language in

language out must be language in the

middle yes and this is an important

misunderstanding yeah can you just

explain that I think that's complete

rubbish yeah

um so

if that were true

and it were just language in the middle

you'd have thought that approach which

is called symbolic AI yeah would have

been really good at doing things like

machine translation which is just taking

English in and producing French art or

something

your thought manipulating symbols was

the right approach for that but actually

neural networks work much better than

Google translate when they switched from

doing that kind of approach to using

your alerts really much better what I

think you've got in the middle is you've

got millions of neurons and some of them

are active and some of them aren't and

that's what's in there

the only place you'll find the symbols

are at the input and at the output

we're not exactly at the University of

Toronto we're close to University of

Toronto at universities here and around

the world we're teaching a lot of people

to code does this still make sense to be

teaching so many people to code

um I don't know the answer to that in

about 2015

I famously said it didn't make sense to

be teaching Radiologists to recognize

things in images

and because within the next five years

um computers will be better at it yeah

are we all about to be Radiologists

though well then Coopers are not better

I was wrong it's going to take 10 years

not five I wasn't wrong in spirit I just

got I factor of two computers are now

comparable with Radiologists at a lot of

medical images yeah they're not way

better at all of them yet but they will

get better yeah so I think there'll be a

while when it's still worth having

coders and I don't know how long that'll

be but we'll need less of them yeah

maybe or we'll need the same number and

they'll be able to achieve a whole lot

more

um was talking about cohere we went over

and visited them yesterday you're an

investor uh in them maybe maybe the

question is just like how they convince

you what was the pitch that convinced

you I want to invest in this so they're

good people

um and I've worked with several of them

yeah and they were one of the first

companies to realize that you need to

take these big language models being

developed to places like Google and

other places open Ai and

um make them available to companies

so there's going to be it's going to be

enormously valuable to companies to be

able to use these big language models

um

and so that's what they've been doing

and they've got a significant lead in

that

so

that's why I think they're going to be

successful another thing you've said

that I just find fascinating so I want

to get you to talk about it is

the idea that there'll be kind of a new

kind of computer that will be

sent to this problem what is that idea

so

there's the biological route to

intelligence where every brain is

different and we have to communicate

Knowledge from one to another by using

language

and there's the current AI version of

neural Nets where you have identical

models running on different computers

and they can actually share the

connection strength so they can share

billions of numbers this is how we make

a bird yeah so they can share all the

connection strengths for recognizing a

bird and one can learn to recognize cats

and the other can learn to recognize

birds and they can share their

connection strengths and now each of

them can do both things right and that's

what's happening in these big language

models they're sharing but that only

works in digital computers because they

have to be able to do identical things

and you can't make different biological

brains behave identically so you can't

share the connections yeah but why

wouldn't we stick with digital computers

because of the power consumption you

need a lot of power it's getting less as

chips get better but

um you need a lot of power to do this to

run a digital computer you have to run

it at such high power that it pays

exactly in the right way whereas if

you're willing to run at much lower

power like the brain is

then

you'll allow a bit of noise and so on

but that particular system will adapt to

the kind of noise in that particular

system and the whole thing will work

even though you're not running it at

such high power that it pays exactly as

you intended

and the difference is the brain runs on

30 Watts a big AI system needs like a

megawatt

so we're training on 30 watts and these

big a systems are using because they've

got lots of copies of the same thing

they're using like a megawatt so you

know you're talking factor of the order

of a thousand in the power requirements

and so I think there's going to be a

phase when we train on digital computers

but once something's trained

we run it on very low Power Systems

so if you want your toaster to be able

to have a conversation with you and you

want to chip in it that only costs a

couple of dollars

but can do chat gbt that'd better be a

low power animal chip what are kind of

like the next

things you think this technology will do

that will impact people's lives

it's hard to pick one thing I think this

it's going to be everywhere right it's

already sort of sort of getting to me

everywhere chat GPT is just made a lot

of people realize it it's going to be

everywhere but it's already you know

when Google does search it uses big

neural Nets to help decide what's the

best thing to show you we're at a

transition point now where

chat gbt is this kind of idiot savant

and it's also doesn't really understand

about truth

is being trained on

lots of inconsistent data it's trying to

predict what someone will say next on

the web yeah and people have different

opinions and it has to have a kind of

blend of all these opinions

so that it can model what anybody might

say

it's very different from a person who

tries to have a consistent world view

yeah particularly if you want to act in

the world

um it's good to have a consistent world

view and I think what's good one thing

that's going to happen is we're going to

move towards systems that

um can understand Different World Views

and can understand that okay if you have

this world view

then this is the answer and if you have

this other world view then that's the

answer we get our own truths well that's

the problem right because what you and I

probably believe unless you're an

extreme relativist is that actually is a

truth to the matter

certainly on many topics on many topics

or even most topics yeah like the Earth

is actually not flat it just looks flat

right

yeah so do we really want a model that

says well for some people like we don't

know that's going to be a big issue and

we don't know we don't know how to deal

with other present yeah and I don't

think Microsoft knows how to deal with

it either they don't and it seems to be

a huge governance challenge who makes

these decisions

it's very tricky things you don't want

some big for-profit company deciding

what's true but they're controlling how

we turn the neurons Google is very

careful not to do that at present

um

what Google will do is refer you to

relevant documents which will have all

sorts of opinions in them well they

haven't released their chat product at

least as we speak right um but we've

seen at least the people that have

released chat products feel like there

are certain things they don't want to be

said by their voice right so they go in

there and meddle with it so it won't say

offensive things yeah but there's a

limit to what you can do that way

there's always going to be things you

didn't think of right yeah so I think

Google is going to be far more careful

than Microsoft when it does release the

chatbot yeah and it'll probably

um come with lots of warnings this is

just a chatbot and and don't necessarily

believe what it says careful in the

labeling or carefullying in the way they

meddle with it so it doesn't do lousy

things all of those things careful in

how they present it as a product and

careful in how they train it yeah um

and do a lot of work to prevent it from

saying bad things and well who gets to

decide what a bad thing is

some bad things are fairly obvious but

many of the most important ones are not

yes so that is a big open issue at

present I think Microsoft was extremely

Brave to release chat GPT yeah do you

see this as like a larger some people

see this as a larger societal thing we

need either regulation or big public

debates about how we handle these issues

well when it comes to the issue of

what's true

I mean do you want the government to

decide what's true

speak problem right yeah you don't want

the government doing it either I'm sure

you've thought deeply on this question

for a long time how do we navigate the

line between you just send it out into

the world and we find ways to curate it

like I say I don't know the answer and I

don't believe anybody really knows how

to handle these issues we're going to

have to learn quite fast how to handle

these issues because

it's a big problem with president but

yeah how how it's going to be done I

don't know but I suspect as a first step

at least these big language models are

going to have to understand that there

are different points of view and that

completions it makes a relative to a

point of view some people are worried

that this could take off very quickly

and we just might not be ready for that

does that concern you

it does a bit until quite recently I

thought it was going to be like 20 to 50

years probably have general purpose AI

yeah and now I think it may be 20 years

or less so okay some people think it

could be like five

is that silly I wouldn't completely rule

that possibility out now and where's pre

a few years ago I would have said no way

okay and then some people say AGI

could be massively dangerous to humanity

because we just don't know what a system

that's so much smarter than us will do

do you share that concern

I do a bit

um

I mean obviously what we need to do

is make this synergistic

have it so it helps people and I think

the main issue here well one of the main

issues is the political systems we have

so I'm not confident that President

Putin is going to use AI in ways to help

people

like even if say the US and Canada and a

bunch of countries say okay we're going

to put these guard rails up then how do

you yeah it's it's

pretty for things like autonomous lethal

weapons

okay we'd like to have something like

Geneva conventions like chemical weapons

people decided they were so nasty they

weren't going to use them except just

occasionally but I mean basically they

don't use them people would love to get

a similar treaty for autonomous lethal

weapons but I don't think there's any

way they're going to get that I think if

Putin had an autonomous lethal weapons

he would use them right away this is

like the most pointed version of the

question and you can just laugh it off

or not answer it if you want but what do

you think the chances are of AI just

wiping out Humanity can we put a number

on that

it's somewhere between

100 percent

okay I think I think it's not

inconceivable

okay

that's all I'll say I think if we're

sensible

we'll try and develop it so that it

doesn't but what worries me is the

political Citizens We're in yeah where

it needs everybody to be sensible

there's a massive political challenge it

seems to me and there's a massive

economic challenge in that you can have

a whole lot of individuals who pursue

the right course and yet the profit

motive of Corporations may not be as

cautious as the individuals who work for

them

Maybe

I mean I only really know about Google

that's the only Corporation I've worked

in and they've been among the most

cautious they're extremely cautious

about AI because they've got this

wonderful search engine that gives you

the answers you want to see and

they can't afford to

risk that yeah whereas Microsoft has

Bingham well if being if being

disappeared in Microsoft would hardly

notice yeah

but it was easy for Google to take it

slow when there wasn't someone nipping

at their heels and this seems to be

exactly yeah so Google has actually been

in the lead I mean Transformers were

invented at Google right the big

language models early ones were at

Google but and they kind of kept it in

your lab they're being much more

conservative and I think it might be so

yes but now they feel this pressure yeah

and so they're trying to they're

developing a system called bad that

they're going to put out there and

they're doing lots and lots of testing

of it

um

but they're going to be I think a lot

more cautious than Microsoft

you mentioned autonomous weapons let me

give you a chance just tell the story

what's the connection between that and

how you ended up in Canada okay there

were several reasons I came to Canada

but one of them was certainly not

wanting to take money from the U.S

defense department this was at the time

of Reagan when they were mining the

harbors in Nicaragua

and it was interesting I was at a big

university in Pittsburgh

and I was one of the few people there

who thought that mining the harbors in

Nicaragua was really wrong

[Music]

um

so I felt like a fish out of water and

you saw that this was where the money

was coming from for this kind of work so

that department almost all lemon came

from the Transformer you started to talk

about the concerns that bringing this

technology to Warfare could present what

what are your concerns

oh that

um the Americans would like to replace

their soldiers by autonomous by AI

soldiers

and they're trying to work towards that

and what evidence do you see of them

I'm on a mailing list

from the U.S defense department I'm not

sure they know I'm on the meeting list

it's a big list they didn't notice

you're there you might be off tomorrow I

might be off tomorrow what's on the list

oh they just describe various things

they're going to do there's some

disgusting things on there okay we'll

discussed you the thing that disgusted

me most was a proposal for a

self-healing minefield

so the idea is look at it from the point

of view of the minefield

when some silly civilian trespasses into

the Minefield they get blown up and that

makes a hole in the poor my field so

it's got a gap in now so it's not fit

for purpose yeah so the ideas may be

nearby Minds could communicate or maybe

they could move over a bit

and they call that healing and it was

just the idea of

talking about healing for these things

that blow the legs off children I mean

and the healing being about the

Minefield healing yeah that disgusted me

there is this argument that though the

autonomous systems might play a role in

helping the Warfighter it's ultimately a

human making the decision here's what

worries me if you wanted to make an

effective autonomous Soldier

you'd need to give it you'd need to give

it the ability to create sub goals

in other words it has to realize things

like okay I want to kill that person

over there but to get over there how am

I going to get over there and then it

has to realize well if I could get to

that road I could get there more quickly

so it has a sub goal of getting to the

road

so as soon as you give it the ability to

create his own sub goals

it's going to become more effective

and so people like Putin are going to

want robots like that and but as soon as

it's got an ability to create sub goals

you have What's called the alignment

problem which is how do you how are you

sure it's not going to create sub goals

that are going to be

um not good for people not good for you

who knows who's on that road who knows

on that road and if these systems are

being developed by the military the idea

of wiring in some rule that says never

hurt a person

well that's they're being designed to

eat at people yeah

do you see any way out of this is it a

treaty is it what is it

I think the best batch is something like

a Geneva Convention but it's going to be

very difficult I think if there was a

lot of public outcry that might persuade

I can imagine the Biden Administration

going for something like that with

enough public outcry

but then you have to deal with Putin

yeah

um okay we've covered so much I think I

have like two more things there's one

more thing I want to say yeah yeah go

for it you can ask me the question some

people say that these big models are

just autocomplete well on some level the

models are autocomplete we're told that

the large language models they're just

predicting the next word is that not so

simple no that's true they are just

predicting the next word and so they're

just auto-complete but

ask yourself the question of what do you

need to understand about what's being

said so far in order to predict the next

word accurately and basically you have

to understand what's being said to bring

language so you're just already complete

too

um in the sentences they are you can

predict the next word maybe not as well

as chat gbt yeah but to do that you have

to understand the sentence so let me

give you a little example from

translation

it's a very Canadian example okay

suppose I take the sentence the trophy

would not fit in the suitcase because it

was too big

and I want to translate that into French

well when I say the trophy would not fit

in the suitcase because it was too big

you assume the it refers to Trophy I do

and in French

trophy has a particular gender so you

know what pronoun to use yeah but

suppose I say the trophy would not fit

in the suitcase because it was too small

now you think that it refers to suitcase

right and that has a different gender in

French

so in order to translate that sentence

to French you have to know when it

wouldn't fit in because it was too big

it's the trophy that's too big and when

it wouldn't fit in because it was too

small it's a suitcase that's too small

and that means you have to understand

about spatial relations and containment

and so on yeah so you have to understand

just to do machine translation or to

predict that pronoun if you want to

predict that pronoun you've got to

understand what's being said it's not

enough just to treat it as a string of

words yeah yeah

I mean this gets me to another thing

you've pointed out which is kind of a

either exciting or troubling idea that

you working intimately in this field for

as long as anyone describe the progress

as well we had this idea and we tried it

and it worked and so we get a couple

decades of back propagation we have this

idea for a Transformer now we'll do some

trick but it could there's hundreds of

other ideas that haven't been tried out

yes so I think even if we didn't have

any new ideas just making computers go

faster and getting more data will make

all this stuff work better we've seen

that as they scale up chat gbt it's not

radically new ideas there I think it's

just more connections and more data to

train it with yeah but in addition to

that there's going to be new ideas like

Transformers and they're going to make

it work much better are we close to the

computers coming up with their own ideas

for improving themselves

um yes we might be and then it could

just go fast that's an issue right we

have to

think hard about how to control that

yeah can we

we don't know we haven't been there yet

but we can try okay that seems kind of

concerning

um yes

do you have any you're seen as like a

Godfather of this industry do you have

any concern about what you've wrought

I do a bit on the other hand

I think

whatever's going to happen is pretty

much inevitable that is one person

stopping doing research wouldn't stop

this happening if my impact is to make

it happen a month earlier

that's about the limit of what one

person can do there's this idea of the

and I'm going to get it wrong the short

Runway and the long takeoff maybe we

need time to prepare or maybe it's

better if it happens quickly because

then people will have urgency around the

issue rather than like creep creep creep

do you have any like thoughts on this I

think time for repair would be good and

so I think it's very reasonable for

people to be worrying about those issues

now even though it's not going to happen

in the next year or two yeah people

should be thinking about those issues we

haven't even touched on job displacement

um which is just my mistake for not

bringing it up

is this just going to eat up just job

after job after job after job

I think it's going to make jobs

different people are going to be doing

the more creative end and less of the

routine end but what's the creative if

it can write the poem and make the movie

and all of that

well if you go back in history and look

at ATMs these cash machines came along

and people said that's the end of bank

tellers it wasn't actually the end of

Bank tell us

um the bank tellers now deal with more

complicated things

and take coders so people say you know

these things can do simple coding and

usually get it right you just need to

get it to write the program and then

just check it so you'll be able to work

10 times as fast

well either you could have

10 of the programmers well you could

have the same number the program as

producing 10 times as much stuff yeah

and I think there's going to be a lot of

trade-offs like that you'll the once

these things start being creative

they'll be hugely more stuff created

this is the biggest technological

advancement sense

is this another Industrial Revolution

what is this how should people think of

it

I think it's comparable in scale with

the Industrial Revolution or electricity

electricity maybe the wheel or maybe the

wheel

yeah

that was earlier yeah okay so buckle up

yeah

one of the reasons I got a Toronto got a

big lead in AI

is because of the

policies of the granting agencies in

Canada which don't have much money but

they use some of that money to support

curiosity-driven basic research okay and

so in the states

the funding comes you have to say what

what products you're going to produce

with it and so on yeah yeah some of the

government money quite a lot of it is

given to

professors to employ graduate students

and other researchers to explore things

they're curious about and if they seem

to be good at that then they get more

money three years later

and that's what supported both Joshua

Benja and me it was money for curing

curiosity driven basic research and

we've seen that before even through

Decades of not being able to show much

yes even through decades not being able

too much so that's one thing that

happened in Canada another thing that

happened was there's a Canadian

organization called the Canadian

Institute for advanced research

that provides extra money to professors

in areas where Canada is good and

provides money for professors to

interact with each other when they're

far apart like in Vancouver and Toronto

but also to interact with researchers in

other parts of the world

um like America and Britain and Israel

and so on and

see far setup of programming AI is set

at one originally in the 1980s which is

the one that brought me to Canada which

was in symbolic AI

yet you came I was an oddball okay I was

kind of weird because I did this stuff

everybody else thought was nonsense they

recognized that I was good at this kind

of nonsense and so if anyone's going to

do the nonsense it might as well be him

one of my letters of recommendation said

that it said you know I don't believe in

this stuff but if you want somebody to

do it Jeff engines to go okay

um and

then after that program finished I went

back to Britain for a few years and then

when I came back to Canada they decided

to fund a program in deep learning

essentially sentience I think you have

complaints with the even just how you

define that right yeah I

when it comes to sentience I'm amazed

that

um people can confidently pronounce

these things are not sentient and when

you ask them what they mean by sentient

they say well they don't really know

so how can you be confident they're not

sentient if you don't know what sentient

means so maybe they are already

who knows I think whether they're

sentient or not depends on what you mean

by sentient so you better Define what

you mean by sentient before you try and

answer the question are they sentient

does it matter what we think or does it

only matter whether it effectively acts

as if it is sentient

it's a very good question Matt

and what's your answer I don't have one

sure okay because if it's not sentient

but it decides for whatever reason that

it believes it is and it needs to

achieve some goal that is contrary to

our interests but it believes in its

interests does it really matter if in

any human I think a good a good context

to think of this in is an autonomous

Lethal Weapon yeah okay so it's all very

well saying it's not sentient but when

it's hunting you down to shoot you

um

yeah you're going to start thinking it's

sentient

we're not really caring not an important

standard anymore the kind of

intelligence we're developing is very

different from our intelligence so it's

this idiot savant kind of intelligence

yes so it's quite possible as

if it is a tool center is essentially in

a somewhat different way from us but

your goal is to make it more like us and

you think we'll get there and my goal is

to understand us oh okay no but yeah and

I think I think the way you understand

us is by building things like us okay so

that's I mean the physics is called

Richard Feynman said you can't you don't

can't understand things unless you can

build them that's the real test of do

you understand it and so you've been

building so I've been building yeah