The Truth about AI 3/3 - 2023 Christmas Lectures with Mike Wooldridge

'The future of AI: dream or a nightmare?' Professor Mike Wooldridge is joined by leading experts to grapple with the future of AI. What opportunities and dangers lie ahead as AI continues to evolve?
This is the third of the 2023 CHRISTMAS LECTURES from the Royal Institution, supported by CGI, on the theme 'The Truth About AI'.
With thanks to contributors Paul Newman (Oxa), Elke Schwarz (Queen Mary University), Maurice Fallon (Oxford Robotics Institute), Sue Black (Durham University), Vian Bakir (University of Bangor), Hassun Ugail (University of Bradford), Kate Devlin (King's College London), Alisa Patotskaya (Immersive Fox), Perla Maiolino (University of Oxford), Stuart Russell (University of California, Berkeley).
Watch the first lecture here: • The Truth about AI 1/3...
Watch the second lecture here: • The Truth about AI 2/3...
In the 2023 CHRISTMAS LECTURES, supported by CGI, Mike Wooldridge will tackle the most important and rapidly evolving field of science today - Artificial Intelligence (AI). Find out more about the lectures here: www.rigb.org/christmas-lectures
Not since the World-Wide Web emerged 30 years ago has a new technology promised to change our world so fundamentally and so swiftly as AI does. Today’s AI tools such as ChatGPT and AlphaGo are just a hint of what is to come. The future of AI is going to be quite a journey, and the 2023 CHRISTMAS LECTURES to be broadcast on BBC Four and iPlayer in late December, will give us a guided tour.
AI has increasingly grabbed the headlines in recent years - generating excitement and concern in equal measure - but what should we believe? In these lectures Mike will reveal how AI works and how it will affect our lives - and will tackle head-on our hopes and fears for this most fascinating of fields.
----
Mike Wooldridge is an academic and author specialising in Artificial Intelligence (AI).
As an academic, he is a professor of AI at the University of Oxford, where he served as Head of Department of Computer Science from 2014 to 2021; he is also Director for AI at The Alan Turing Institute in London. He has received multiple awards for both research and education, including in 2020, the Lovelace Medal from the British Computer Society - the leading award for a UK computer scientist - and in 2021 the Outstanding Educator Award from the Association for the Advancement of AI (AAAI).
From 2014-16, Mike was President of the European Association for AI, and from 2015-17 he was President of the International Joint Conference on AI (IJCAI). He is currently Editor-in-Chief of Artificial Intelligence, the leading journal for AI, established more than 50 years ago and has been invited to give evidence on matters relating to AI to multiple government committees.
As an author, he has written nine books, which have been translated nine times. His books include two popular science introductions to AI: the Ladybird Expert Guide to AI (Penguin, 2018) in the iconic British book series, and The Road to Conscious Machines (Pelican, 2020).
Mike gives frequent public lectures on AI, including at the Hay Festival and Cheltenham Science Festival and is regularly interviewed by the media.
----
Join this channel to get access to perks:
/ @theroyalinstitution
Subscribe for regular science videos: bit.ly/RiSubscRibe
The Ri is on Twitter: / ri_science
and Facebook: / royalinstitution
and TikTok: / ri_science
Listen to the Ri podcast: podcasters.spotify.com/pod/sh...
Donate to the RI and help us bring you more lectures: www.rigb.org/support-us/donat...
Our editorial policy: www.rigb.org/editing-ri-talks...
Subscribe for the latest science videos: bit.ly/RiNewsletter
Product links on this page may be affiliate links which means it won't cost you any extra but we may earn a small commission if you decide to purchase through the link.
Join this channel to get access to perks:
/ @theroyalinstitution
Subscribe for regular science videos: bit.ly/RiSubscRibe
The Ri is on Twitter: / ri_science
and Facebook: / royalinstitution
and TikTok: / ri_science
Listen to the Ri podcast: podcasters.spotify.com/pod/sh...
Donate to the RI and help us bring you more lectures: www.rigb.org/support-us/donat...
Our editorial policy: www.rigb.org/editing-ri-talks...
Subscribe for the latest science videos: bit.ly/RiNewsletter
Product links on this page may be affiliate links which means it won't cost you any extra but we may earn a small commission if you decide to purchase through the link.

[@kh_167]

Petition to release the archives of all the other Ri Christmas Lectures as well👍

[@Alice_Fumo]

This is the best lecture for children I have ever seen. A lot of very difficult topics have been broken down excellently in ways children can understand them and everyone seemed quite interested all the way through despite the length of this lecture. Truly outstanding.

[@mand6029]

the future of AI is absolutely fascinating , frankly speaking, thank you for releasing this video to the public

[@brianstarr]

I found it interesting how the children were so well behaved and generally interested. I doubt you could maintain order here in the states. I also liked the wide age range of children. Good job RI, you covered the key points. I hope the audience remembers this lecture and is inspired to ask more questions about technology. Soon hardware that simulates the human brain will be in autonomous robots. Now is the time to jump into technology.

[@FaitCeQueVeut]

Young ones are that much attentive at the World Science Festival.

[@vijaymujumdar5617]

An outstanding presentation. Captivating for both, young and old. Truly an example of excellence. Please keep up the good work.

[@BrilliantOverlord]

You started calling it SI (Simulated Intelligence) in the Turing Lectures. Much better description imo. Does a better job of capturing the idea of what is really going on imo.

[@pklausspk]

My hopes are rising while seeing these interested young people. It is getting dark outside.

[@RaghavIyer]

Amazing! Really Interesting

[@socialtraffichq5067]

That's the problem with an AI enabled self-driving car... It's driving on the wrong side of the road!

[@MelroyvandenBerg]

AI will solve this eventually in the UK...

[@keep-ukraine-free528]

All three lectures in this series were exceptional, and all three will be a treasure for generations. Suitably, this third was softened for children -- since it dealt with threats & futures. I agree with not scaring young children with the panoply of threats on the horizon (complex threats from ASI - "Super AGI"). The lecture nudged right up to the idea of ASI, without decloaking it. Adult viewers may take & extrapolate the hints. Superlative info for children (tomorrow's researchers).

[@haniamritdas4725]

The children's stories of ages past were warnings. This is a lull to sleep. Not helpful to the next generation. Just this one.

[@keep-ukraine-free528]

@@haniamritdas4725 You made a valid point. My initial reaction was similar to yours. But after thinking about it, I considered the possibility, "if we don't address the serious risks correctly, will there even be a next generation?" Only if today's young adults & youth manage & solve this, will there be future generations of thinking children. We must help today's youth to fully understand the complexity, to not overreact, and definitely to not under-react. They need an even-keeled unemotional mindset about many potential coming threats. Scaring them while they're children will make them emotionally reactive -- that's counterproductive.

[@thomasAnders2765]

I love this series! Fabulously done, and even understand most of it (German); It's a bit like an English lesson for me ;-) I would advise Mike to speak more calmly, sometimes you get the impression he's on the verge of screaming, which is a problem for me as a disabled person. He explains it so well and the inserts are so well done - wonderful.

[@TheCommuted]

I have to expect the future of AI is like the corporate veil. A way of denial of responsibilities for harm and negligence. When the AI regression finds leaving information off of a label sells product the AI inherent lack of morality will become its formal utility.

[@mrvzhao]

Inorganic stuff ->organic stuff ->...->single cell->multicellular ->not so smart creatures ->smarter creatures ->thinking machines -> ... -> who knows?! We are all just part of the universe's bootstrapping process; not sure if AI's will be the final stage, but we are certainly not - whether we like it or not

[@thehybridtech3260]

37:00 best example on AI

[@anthonyshiels9273]

Boston Dynamics entry level robot Spot can execute preprogrammed tasks. Their enterprise model has autonomous capabilities like the one at Starbase.

[@charlottelanvin7095]

remind what Starbase is again?

[@anthonyshiels9273]

​@@charlottelanvin7095 The device that you have typed that comment into is a computer that comes with a Search Engine. Why don't you use this facility to look up "starbase"? Please note that this is a Science page...

[@skierpage]

​@@charlottelanvin7095Starbase is SpaceX's facility in Boca Chica Texas where it builds, tests, and launches its next-generation Starship fully reusable rocket. After one crash, SpaceX released footage of its Spot robot coming out to survey the damage. The problem for Boston Dynamics is, a cheaper drone could have done it just as well.

[@enterpassword3313]

​@@charlottelanvin7095its where Darth Vader lives

[@MelroyvandenBerg]

AI doesn't have feelings.. yet... but since you said it yourself, we also do not understand our own consciousness, so it's a matter of time before AI and robots do have those emotions eventually.. since we do not understand ourselves and we also do not understand the internal neural networks.

[@vansf3433]

AI obviously can help humans understand and apply their invented notions better to the real physical world. But it doesn't mean that AI should be deployed in all human aspects of life. No software can be perfect, which means that software errors for one reason or another can cause disastrous consequences, without human supervision. Additionally, such ideas can cause massive human unemployment if there is no law for such AI development. Currently, no country in the world has got any specific law for AI development to prevent bad human purposes

[@enterpassword3313]

Good, ideal unemployment rate is 100%

[@PedroAmA]

Wow but still scary I wana have control

[@davehall8584]

Awesome lecture!

[@DrDeuteron]

It's pretty easy to figure out if you're being exposed to disinformation...eg., you're at an Ivy League school, you're watching cable news, reading the New York Times, and so on.

[@keep-ukraine-free528]

@DrDeuteron Clearly some people failed to understand the point in the lecture -- that disinformation will be spread by AI and by people fooled by AI. Your comment shows your human weaknesses. AI will easily exploit you -- to spread disinformation. Congrats!

[@uncletungsten5253]

Well said!

[@Bob-qq4is]

Conservative media is 10 times the disinformation of mainstream media. People who watch pbs news hour have a much more accurate view of the world then people who only listen to independent media and Facebook memes

[@richard_d_bird]

ai will free us from the drudgery of our miserable jobs, and then watch impassively as we shrivel up and die unemployed and freezing. but will it laugh on the inside? we may never know

[@pkluckyindia]

👏👏👏👏

[@gidi1899]

hack the car LA system, change the training result confirmation to 'try to crush', and after learing mistakes (not crushing is a mistake), update all the other non hacked cars, Do you really want to give this option to someone?!

[@antonioa8869]

Why do humans believe that AIs will behave with others as humans do?

[@Whyiseverythingthesame]

I didn't think you should hit the toaster but that's mainly because it's not a printer.

[@YTantirungrotechai]

great lectures as always. Thank you RI !

[@anthonyshiels9273]

Re Deep Fakes: Facebook and KZbin are featuring ads where Elon Musk is promoting cryptocurrency. It looks and sounds like Elon but the entire video presentation was generated by Artificial Intelligence.

[@enterpassword3313]

Lol its the most obvious fake. Idk how anyone can fall for that scam, whats worse is youtube doesnt care and they keep showing the scam ads.

[@TazzSmk]

I don't think it's so simple to say AI has no consciousness, emotions etc... because in basically same way an actor performs in a movie, AI has the knowledge/data to pretend consciousness, emotions etc... and even it's been proven AI can provide more accurate replies/responses when you give it a role; so the question is, if people can't see through chatbot AI giving false information, how can they tell if it has consciousness or emotions? in a way people can't often even distinguish physical limitations of mental disorders for example?

[@johncarter1150]

Hey kids here's your future... Isn't it complicated, sad.... Any questions?

[@ratti80]

AI is choosing the targets for the IDF. So too late.

[@johncarter1150]

You'll have to try to live indigenous for autonomy.

[@Sonsequence]

The other thing people forget when they think about the risk of super intelligent AI is that it will already be helping at the executive level in large companies, will be able to steal huge sums of money. Money means you can pay people to do what you lack the physical limbs for. And once you've gained control you don't need to coerce some through greed, you can control all through fear. In a military dictarship, you swear loyalty to the leader because the personal consequences of not doing so are unthinkable. The only way to contain a strong AI with such intentions is to not use it. That's why the research focus is on aligning its intentions with ours', not on containment of its capacities.

[@enterpassword3313]

You will own nothing, you will be happy

[@cardrivingdude]

So when did it pass the Turing test? Oh, it hasn't? Oh OK.

[@brockstanford7608]

Turing was a genius, and his prediction about AI in the year 2000 was spot on. The turing test has been bested.

[@skierpage]

The Turing test was passed years ago without the benefit of large language models, by having a primitive AI convincingly act like a bored distracted teenager. It was a great thought experiment, but it's no longer relevant. Right now, if you could text chat with an average human being or ChatGPT, it is obvious which one you choose to answer a college-level test question, improve your writing, develop a legal strategy, write a program for you, etc. AGI in language has already arrived!

[@cardrivingdude]

​@@skierpage All that was generously taken into consideration before I left my comment. AI does not yet exist in the public sector, and my guess is that it does not yet exist in the private sector either.

[@cardrivingdude]

@@brockstanford7608 That's up for debate. So, no it hasn't.

[@stephmaccormick3195]

With AI the problem is in liability. You kill a pedestrian with a autonomous car, or hurt an infant with Spot... who's responsible? You, software?

[@RajputAbdulWaheedBhatti]

Very impressive

[@socialtraffichq5067]

Let me just give you a spoiler alert. And the not too distant future government is not going to let human drivers on the road.

[@uncletungsten5253]

We will rise before then!☝🏻

[@anthonyshiels9273]

Children today will not need to learn how to drive a car.

[@k1ry4n]

​@@uncletungsten5253We will stop you before.

[@ThatRainbow]

19:07 Bellybutton lol

[@alexijohansen]

I always run far away from people who claim to know “The Truth”.

[@k1ry4n]

Where did you see such people?

[@enterpassword3313]

​@@k1ry4nthe religious... "i trust the science" people... political pundits... self help gurus

[@kencory2476]

Mike is incredibly tall compared to his guests.

[@Sonsequence]

The fact this quietist take on AI was chosen as the Christmas lecture will become one of those oft passed around factlets used to make a point about how backward and confused people were in the olden days.

[@socialtraffichq5067]

Is this a video for children

[@uncletungsten5253]

These extreme leftists indoctrinate our children with their BLM, CRT, sick trans-activist, antisemitic, and now normalising AI agenda. Ri, you’ve been exposed. ☝🏻

[@johncarter1150]

​@@uncletungsten5253extremophile, you have been canceled.

[@anthonyshiels9273]

Royal Institution lectures are student and child friendly. Look up their Mission Statement ...

[@noneofyourbizness]

when considering the presentation/delivery of existential threats to our kids, this is a very different approach than that of the approach to climate change, for example.

[@MelroyvandenBerg]

the end result might be the same lol lol

[@richard169]

Horrifyingly upbeat. Watching the faces of the children, I cringe at our lack of preparation.

[@mymysticalside2842]

You are all absolute nutters!

[@onecalledchuck1664]

The more people injured and unalived by driverless AI cars now, the fewer people injured and unalived accidentally by driverless AI cars in some distant future. We're all just Elom Nusk's unpaid and unasked beta testers on the road.

[@starhacker6411]

💀 wtf is this

[@skierpage]

There's good evidence that Cruise and Waymo's robotaxis are already safer than human drivers. As a pedestrian or fellow road user, you should want to be surrounded by car drivers that are never distracted, are always scanning their surroundings in all directions, always use their turn signals, and obey traffic laws. Tesla so far overpromises and underdelivers. Anyone who turns on its so-called Full Self Driving has to agree to a message saying they need to keep their hands on the wheel and be prepared to take over at any moment. It's much less clear that today such shared responsibility is safer overall.

[@enterpassword3313]

Its already a better driver than most people i know lol.

[@thedude8046]

Stop normalising AI!

[@Hecarim420]

ABOMINABLE HERETIC INTELLIGENCE MUST BE PURGED

[@darylallen2485]

🤖

[@johncarter1150]

Sad new artificial world.

[@Vadjong]

It already is.

[@eoinokeeffe7014]

No!

[@danamulter]

Oh look, more farming heartstrings with children while deliberately obfuscating threats so RI can make money. Unsubscribed. The RI can no longer be trusted.

[@jessasna3532]

There is a part in there about biased opinions, but you wouldn't know, because you clearly haven't watched this video

[@DrDeuteron]

You misgendered Dido the Dog as non-binary.

[@fanofhifi]

Really? This is your platform??

[@MelroyvandenBerg]

@@fanofhifi I laughed .. I hope this person was not serious

[@uncletungsten5253]

Lol!!!

[@haniamritdas4725]

This is the tenor and format of a children's clown show. The end of science is artificial idiocy 🥸🤡

[@TazzSmk]

I don't think it's so simple to say AI has no consciousness, emotions etc... because in basically same way an actor performs in a movie, AI has the knowledge/data to pretend consciousness, emotions etc... and even it's been proven AI can provide more accurate replies/responses when you give it a role; so the question is, if people can't see through chatbot AI giving false information, how can they tell if it has consciousness or emotions? in a way people can't often even distinguish physical limitations of mental disorders for example?