Rise of AI Conference 2017 · Berlin

Artificial Intelligence and the Future of Humanity

At Rise of AI 2017 in Berlin, Fabian Westerheide, founder of the conference, gives the opening keynote from the venue hosting a rebuild of Konrad Zuse's Z1, the first programmable computer built in Europe. He recounts how Rise of AI grew from a 15-person meetup at the c-base hackerspace to over 300 attendees, and defines artificial intelligence as algorithms that use data to make and improve their own decisions within given frameworks. He walks through AI already in use, including IBM's Deep Blue and Watson, AlphaGo's win over Lee Sedol, Cortana, Amazon Alexa, self-driving cars, AI-run stock markets, and younger AI companies such as Berlin's Parlamind and Micropsi Industries and London's Exlo Dynamics. He then looks ahead to brain-computer interfaces and an emerging internet of agents by around 2025, describes new jobs such as AI trainer and AI controller, and closes by arguing that without a strong European AI ecosystem, AI will end up thinking American or Chinese rather than reflecting European values.

Key takeaways

  • Rise of AI grew from a 15-person meetup at Berlin's c-base hackerspace in stages to over 300 attendees with a waiting list by 2017.
  • AlphaGo beat Lee Sedol at Go just one year after Nick Bostrom had estimated in Superintelligence that it would take humanity another decade.
  • Berlin's Parlamind and Micropsi Industries, and London's Exlo Dynamics, were building AI for email response, robot learning, and drone swarm control respectively.
  • Westerheide expects new jobs such as AI trainer, AI controller, world creator and community manager as AI takes over more routine work.
  • Westerheide argues that without a strong European AI ecosystem, AI will end up reflecting American or Chinese values instead of European ones.

Detailed notes

From Zuse's Z1 to Rise of AI

Westerheide opens by noting the venue also houses a rebuild of Konrad Zuse's Z1, the first programmable computer built in Europe in the 1930s and 40s, which had 200 bytes of working memory and 5.3 hertz of processing power. He recounts starting Rise of AI as a small meetup at Berlin's c-base hackerspace with 15 people discussing whether a singularity would happen, growing to 70, then 180, and now over 300 attendees with a waiting list, which he reads as AI having reached a critical mass of public interest.

Defining artificial intelligence

Because language is imprecise, Westerheide defines AI as a set of algorithms that use and generate data to make their own decisions and improve within a given framework. He compares algorithms to cooking recipes, and frameworks to the rules and motivation that guide employees or citizens: AI systems process input, use memory, make decisions, and learn by improving, all within rules that humans set for them.

AI already among us

He surveys existing AI systems: IBM's Deep Blue beat the world chess champion in 1996 and 1997, and Google's AlphaGo beat Lee Sedol at Go, a game Nick Bostrom had estimated in his 2014 book Superintelligence would take humanity another ten years to solve, achieved one year later. IBM Watson, using 90 servers with 16 terabytes of memory, won Jeopardy and now runs as a platform for thousands of startup applications. He points to Cortana and Amazon Alexa as everyday AI, self-driving cars that share collective learning across a fleet, AI-run stock markets, and Google's shift to an AI-first strategy across Maps, AdWords and Search.

Younger AI companies from Berlin and beyond

Westerheide highlights younger companies as proof of European AI talent: Parlamind, a Berlin startup that drafts and can auto-send email responses for customer support; Exlo Dynamics in London, which built mission-control software letting a single operator manage large drone swarms by setting premises rather than piloting each drone manually; and Micropsi Industries, based in Berlin, which builds AI that lets robots learn and adapt to a process instead of being explicitly programmed, which he cites to show Berlin's own innovation talent.

Brain-computer interfaces and the internet of agents

He describes existing brain-computer interface headsets that read electrical signals to control computers or let paralyzed people type, and cites Elon Musk's Neuralink and a Facebook research device, expected around 2019, aimed at letting people type by thought, developed by a researcher formerly at DARPA. Looking to around 2025, he predicts an internet of agents in which every app becomes an AI agent that fulfills a person's needs on command, first by typing, later by thought, and describes AI extending into virtual, augmented and physical worlds, including satellite-guided agricultural machinery.

Future jobs and the case for European AI

Westerheide argues AI will not replace people outright but let companies produce the same output with fewer people, and describes emerging roles: AI trainers who teach systems how a job should be done, AI controllers who oversee large numbers of automated systems, world creators building virtual and augmented realities, community managers, and consultants who explain change to others. He closes by warning that without a strong European AI ecosystem, AI trained mostly in the US or China will carry American or Chinese values rather than European ones, which he says is the reason Rise of AI exists.

Quotes

Artificial intelligence is for me a set of algorithms which is using generating data to make their own decisions, to improve within given frameworks.

Fabian Westerheide

We don't have a strong European ecosystem for artificial intelligence, that is why we host this conference.

Fabian Westerheide

Artificial intelligence is just a tool for us to be more human.

Fabian Westerheide

Questions answered in this talk

How does Fabian Westerheide define artificial intelligence?

According to Fabian Westerheide at Rise of AI 2017, artificial intelligence is a set of algorithms that use and generate data to make their own decisions and improve within a given framework of rules and motivation set by people.

How did AlphaGo's win compare to expert predictions?

According to Fabian Westerheide at Rise of AI 2017, Nick Bostrom had estimated in his 2014 book Superintelligence that it would take humanity another ten years to solve the game of Go, but Google's AlphaGo beat Lee Sedol just one year later.

What new jobs does Fabian Westerheide expect AI to create?

According to Fabian Westerheide at Rise of AI 2017, expected new roles include AI trainers who teach systems how a job is done, AI controllers who oversee automated systems, world creators building virtual and augmented realities, and community managers.

Mentioned

Z1 computer, Deep Blue, AlphaGo, IBM Watson, Cortana, brain-computer interface, internet of agents, AI trainer, AI controller

Keep exploring

More from Fabian Westerheide

More from Rise of AI Conference 2017

More on AGI & Future of Intelligence

Transcript

Auto-generated transcript, not edited. · 4,916 words

Show the full transcript

I'm very happy um to have you all here to talk about artificial intelligence and the future of humanity um we start with the past this place here is hosting the original rebuild so far of the Z read by Conrad Souza so it's in the other building and this is the first computer programable computer ever built in Europe so this is a historical place we are speaking here about the future at the place where you can see the achievements in Technologies by hum kind we have done that's why we picked this place that's why we thought this is a perfect place to speak because we are writing history today too we have you know bringing advances um this machine programmed by by tu the 1930s and 40s destroyed during the second world war had 200 bytes of working memory so 200 bytes keep this in mind and 5.3 Herz of processing Powers so everyone who has a a smartphone today knows how damn slow this is very expensive very big first programm computer made in Germany

we started Rise of AI very small and I started a Meetup Group because I wanted to speak speak a couple of years ago with other people about artificial intelligence and um we met in seab base and for those from Berlin they know it everyone else cabase is the first hacker space which we had and it's um spaceship and we were 15 people discussing for a whole evening if there would be a singularity one day or not and they asked me if I could do this a little bit bigger so next time we met and we were 70 people and then we were 180 people and today we are over 300 people so the topic of artificial intelligence finally reached a certain critical mass of interested people and you are all these early adopters who are thinking about this and you the lucky ones who are sitting here because we have a long waiting list because we made a cut a couple of weeks ago said no these who have a ticket they can come no

one else so what is artificial intelligence it's a very tricky questions and whoever you will ask you will get a lot of answers about this words are a very terrible uh way of expressing something uh words is a form of compressing complex thoughts so I try to write down what I how I would describe artificial intelligence artificial intelligence is for me a set of algorithm which is using generating data to make their own decisions to improve within given Frameworks and as I said language is terrible so I need to explain my words with another words and let's start algorithms are nothing else that experience experienced formed in mathematical models so everyone who of you has ever used a cooking receipt a receipt is like an algorithm it gives you a certain step and orders to have a result at the end and we humans we have algorithms too because we have experien and we learned something we we will use them in future then these algorithm they using input they process the input they have

a memory function today long-term memory shortterm memory so they remember things and then they make their own decis decisions what we humans would call thinking yes not thinking thinking but still thinking because they use data to come up with own decisions and then they improve and what we call improving is learning yes if you improve you learn something you did something wrong or right and you do it better next time and within given framework is nothing else than having rules and motivation something everyone knows if you have employees they have to be motivated too they need to have rules and everyone who citizen of the state knows we have rules and laws so the same we do for machines we give them laws we give them rules so they can act on their own within this and the beauty about this is this is adaptable the better they learn the better they make decisions and the larger the framework is the more they will act so words are terrible to explain something so please create

a picture of yourself what is artificial intelligence for you what is the first thought which comes in your mind when you think about AI is it this one I hope not by the way this is Hall 9000 and this is The Stereotype of a dystopian AI or rough AI which starts to kill humans something we don't maybe it's this one it's a little bit more recent it has a female character by it it's from exmachina by the way rough AI too killing humans not good so for me AI is also this one this is IBM de Bloom this this is the computer which won the World Championship in chess or was beating the world champion in chess 1996 and again 1997 so this was an AI this was an AI because the first time or for another time we humans learned we are not perfect machines can beat us in a task because chess was a very human game so far where we thought we only can solve it by the way no animal is

able to play chess and win against us but machines which we buildt was able so this is a computer today it's part of a museum so this is what we do with AI if we don't need it anymore we turn them off we put them somewhere else this is in AI too this is Alpha go and it won against lisod dool in go so whoever has played go before knows go is different than chess it's a game with where you need something like intuition um it is not BR Brute Force so you can't do by math all the possible solutions it's you have to develop a feeling for it and it's a component of strategy still still the AI won it so this is an artificial intelligence too which you don't see but it won a game and still it's a game so it's a fixed framework but within this framework the AI was able to solve it by the way Nick Bostrom wrote in his book super intelligence which is an interesting book in

2014 that humankind needs another 10 years to solve the game of Go one year later Google solved it so whatever even if expert thinks it's not possible one year later you could be wrong it solvable this is in AI too this is IBM Watson most of us have heard about it and just in contrast to the machine which is in the next building yes we spoke about this this has 90 servers together 16 terabyte of memory and each server has eight Quadro core processors with 3.5 GHz so this is over a million times more powerful in the last 80 years maybe cost the same fortune I tell you in 10 years or 20 years every one of you has this in your pocket in the same just smaller this will be in your pocket soon not next year but in the next 10 to 20 years the interesting part about IBM Watson is this is a very strong AI we know it was um winning Jeopardy a very human task too because you know being

a show and so for and it won by lenslight like 2,000 points to 10,000 to 2,000 points like wonderful but also it's more than an AI because a lot of startups and companies use Watson to build their own AI applications so this is a platform AI it's not one AI we speak here about thousands of AI which are running on these systems so one AI is producing more AIS out there um everyone of you has an AI with him yeah in your pocket if you're an Apple user even it's not the most smartest one it's still a form of AI um for the windows users amongst us you can use Kotana um this is a very beautiful piece of AI Amazon Alexa because this is the first AI which is in your home and your family starts interacting with you use this computer not to type you give comments by voice and this is the future too in future you will not type with your computers the next step is you will talk to your

computers because we can talk way faster than we can type typing is super slow talking is way faster and you can hear it when I speed up too much yes so this is a beautiful too because you say what is the weather please order me food um um please order me uh a taxi but please play my favorite Spotify list please turn on the TV make the so we use AI to interact with our environment this AI helps us to control the less sophisticated systems out there self-driving cars each of them if it's di Audi if it's Tesla it's Google they have their own form of AI and this is a very very beautiful example of about the collective intelligence if one Tesla does not know what a white truck is on the street it's tragic but afterwards every Tesla out there knows how to recognize a white truck so so hundred thousands of cars are there learned from one mistake one system was doing with every mile a car is driving if they do

something right or wrong every other car learns so they get better every day significantly this system updates instantly we humans we all know it you had to go to driving schools you had to take lessons you had to learn all these paperworks and it didn't you know even if you know it well all your friends they had to do it too we could not share this knowledge it was really really slow really painful and we are still shitty drivers as you heard in the video before we kill 40,000 people just in the US every year we are terrible at this so that's why we have ai to do the stuff which we are not good in and which we don't want to and which we're super slow well um stock market is mainly run by AI today I don't say anymore you beat the market you beat the bot which you can't it's faster it's more reliable and it is rational we all know that stock markets go down because people are irrational they sell

and buy when they're not supposed to do because it doesn't make sense and they are they are less hard to influence yes um mainly our wealth is today driven by AI because they already run the stock markets and everyone who has stocks and they were going up for the last six years we benefit from this Google is an AI company Google changed their strategy to AI first a couple of month ago it's a full AI company developing if you use Google Maps well you have seen the AI it knows which path is the fastest it knows where the traffic jems is it is you know I I still grew up with using maps and whoever was doing the maps they didn't know the traffic jam they didn't know which route is fast especially outside of the country so there is an AI Google AdWords is an AI because it's optimizing your bits Google search is an AI because it knows what you type in before you type it in and even if you don't know

what you type in it will give you the right recommendations and just imagine you had to go to a library get a book read for the sources get the next book and so for you do this within 2 minutes on Google so this is an AI and there are younger AIS too um this for example is the company parine based here in Berlin and they are able to answer emails if you use Zend this or Gmail for example an email is coming in and the customer support agent sees the POS possible answer already there he can say I want that the AI is answering the email all the time or only if I accept this but this system is able to read your emails and we all hate emails we all I would say get 80% not the spam the spam is already faded out I call it soft spam people who want something from you and you know you don't want to answer it you could use an employee or something else to do

this for you well this is your digital employee and in the future every one of us will have this it's like your smart inbox and you will decide who will get instant responses who not and so far another form of AI is the company exlo Dynamics based in London they have AEM mission control for drones so if you want to control one drone oh that's it's hard but it's possible if you want to do two or three it's hard have you ever tried to control 100 drones at the same time manually impossible this is a task we humans can't do but they have developed the software for this so they have in AI which you only tell well this is the premise please please protect it or this is the premise please surveillance or this is a mesh Network we want to establish and this is the moving Fleet out there and all these drones they have a very minor AI out there and this minor AI is controlled by the larger Ai and the

AI behind it is the human who's giving the task so they use Ai and AI to do and control the machines all the company microps Industries based here in Berlin too I mentioned this to show that we in Berlin have Innovation talents yeah um is teaching robots to learn so they have a software they have an AI which you integrate in a robot and the robot is able to adapt to processes instead of programming your robot the robot will learn from you so artificial intelligence is eating our human world that's a fact why because we want this we all you invest all your time here to learn about this maybe you already have invested money and more time to build an AI or to buy an AI we invest a lot of time to research to develop AIS we invest as investors a company in companies who build AIS billions of dollars go in this we want this we have done this for the last centuries we used machines because we want to work less

we use machines to do the stuff which we don't like to do anymore the same will be about AI just this time it's not about physical strengths it's about your mental strengths so how could artificial intelligence be 2015 because we have enough um speakers today who will tell you how amazing AI is today and how they solved it and so forth so I will give you a little Glimpse outside let's imagine you are hungry uh if not there's Breakfast outside and later we have very good food trucks there so don't start to eat something now but let's say you're hungry if you're hungry today you buy something to you have to go to restaurant you have you can cook you can take out your Lea head app and order sushi your pizza imagine you are hungry you think it and half an hour later food is ready imagine you think you're hungry and the food arrives at your doorstep that's all you have to do do you think this is science fiction or magic the

thing is it's already present this is a very ugly helmet this is a brain computer interface it is able to read the electrical signals in your mind it's not in your mind it's certainly it's in the outside and today you already can control and this technology is old it's for the last 20 years around you can use to control computers you can play this game say go this and this direction um people with Al Als who are totally paralyzed they can use it to type messages so it's mostly used into um uh into medical devices but this works it's just ugly it is slow it's not scalable but it's a technology so you can think and something acts on your computer well then we have a guy called Elan musk and we all know El musk is a little bit crazy overambitious Visionary and he's right at the end so Elan musk just started a company called norer link and he wants to build exactly this kind of device noral link is developing ultra high

band with brain machine interes to connect humans and computers so they are doing this so it's not science fiction anymore and then if you have watched Facebook f8 develop of conference a couple of weeks ago we had this lady there and she spoke about that by 2019 Facebook will bring out a device which should let you type words to a computer just by thinking so in two years maybe three years because always takes longer um or four years you will have a device out there which let you read your thoughts and print them on the computer screen by the way this lady used to be the head of research at darar so darar is finally ing everything this Dara is the American Military agency why do they do it well the governments fear one thing the most and that's their own citizens so that's the way access into your brain there will be Solutions against this but still you have to be aware that this will happen so the first thing which happened with this

is you don't need this device anymore you don't need these devices anymore you don't need to type in because in the next five let's say by 2025 you sitting in front of a device and you think and you will see it imagine that every app you just think and the app opens you say go down and something happens or even more sophisticated imagine you think something and you think so much and it's a feeling did you ever have this moment that you had an argument with your wife or your husband and you didn't understand them happens to me quite often so I wish I would get the feeling what she has because words are so terrible in expressing what you really feel so in future you will be able to share feelings it won't be that you force your feelings on it you will have your own permission you will accept what you have like the same like words you only speak if you want to and if you don't want to speak nothing will

happen so the same will be what but you will be able to share it I like public speaking and I'm not perfect I speak too fast I like to walk around I sometimes I'm over the time so I had to practice as a kid I was as a kid always sitting first row when the teacher asked me who wants to give a talk I was like I do it um so I was the nerd and uh NS Rule the World by the way today yes everyone knows who has an IT department or technical unit and I I practice a lot I practice hours of it but I'm far from perfect so I'm pretty sure someone in this room is better than me just imagine you can share your skills with me everyone can share his skills with me you will come to a point where you know something and you want to share it you share not only your words or you give a workshop about this you can share your experiences your algorithm in

your brain plus your storage your data set you can share it free one time so then this is real just without the actly plug in your head it will be just Wireless we will come what I call is the internet of agents and this is coming back to AI so how will you communicate with AI in seven years or eight years well you will have them in your mind in a certain way and you just communicate say oh please order food there or please um read the emails for me or please make an appointment uh for dinner with my wife tonight or please send my mom mother some flowers to they use apps for this or websites whatever what is a website or an app will be a bought and everything which will be bought will be an agent who's fulfilling your purposes your needs AI is serving to these needs you have and you will have these orders and today you type these orders and at this time you will think it well AI

will run our world by then it's already running today it will run our Digital World yes we you can't create virtual realities and all these worlds without the help of AI and you can't create an augmented world which is coming out without the help of AI it will help you to filter and by the way Facebook just published a new tool which is creating augmented pictures so you running around today like Snapchat filters just more with everything else in future you will see that when you use Facebook all your environment will be labeled you will see you out there who liked what through your lenses because they have published this so it will come too in the next couple of years and AI will run our physical word just take this example an AI sitting in a satellite the satellite is screening the field it's deciding when the machine is supposed to go out and the machines uh are doing the job today already less than 2% are uh working for uh agriculture it used

to be 98% why because we use machines these machines just will be smarter so do we want this I see skeptical faces I know this and I don't speak about something which will happen I don't speak about something which I want I just say this is what I see which will come so whatever things happens there will be challenges there will be obstacles there will be setbacks there will be things we don't like but we can't change it that's why we're here to learn about this we need to adapt we need to change as a society as governments as corporations as individuals with every technology environment will be different and we have to adapt to this so normally I people ask me well AI is eating the drops and you saw this in the video and it's true I don't say this won't happen but the thing is AI will not replace you right away everything you do today it's safe it's more that you will use AI yourself to be more productive in your

job which means on the midterm your company just needs to hire less people to get the same output hopefully the time you're saving you can spend with your family and not working even more so today I just brought you as a a glimpse some jobs which will be used for AI in the future one thing is AI trainer today it's a very technical job but it will be less and less technical we need people out there who are teaching a how the world looks you're teaching your kids something we'd send them to school you have a training on the job you you go to university whenever you learn something AI can learn it too so we need people out there who are training AI how their job is supposed to be done how they can help them how to see this we have to teach them the Frameworks the rules the motivation the ethical values all this needs to be trained you don't program AI anymore you teach AI you train AI well then we

need AI controllers and this will be a fun job if you have ever played Farm ball or Starcraft or something like this it will be similar so why should you go out on the field and drive the machine why don't you have a dashboard where you can see all the machines there and you optimize there why because we are the intelligence behind it we use AIS to do stuff which you can't do at this moment well we will have World creators especially with the augmented reality and the virtual reality we need a lot of people who can fulfill their dreams who who want to be God and create new worlds for others to fulfill their fantasies whatever they want to do we need this and this will be a very cool job by the way everyone will be able to create a matrix for their own everyone will create Matrix for others there won't be one Matrix we all enter it's more that everyone can opt in as he wants to because we create these

worlds with the help of AI Community managers the more we use technology the more people are important that's why you're here physical that's why you don't have a streaming pass that's why you don't wait for the YouTube video you're here because you want to meet people you want to connect you want to get this energy that's why people are important but people will be important in the virtual and in the offline world too and yes consultancy we need them too they we won't don't get rid of them sorry for those yes but I did it too before um they will be there because with every change you need people explaining change to those who do not understand it so lawyers consultancy everything which is a people business the people power this will stay this is the emotional part this which which we can't solve because we don't have understand it ourself yet so everything which is based on people this will stay everything which is based on a computer you're sitting front of computer and

putting data in it's getting data out you don't want to do this I don't like sitting in front of the computer I like speaking with you people so next question what values should the AI have because if we go like we go today it either will think Chinese or we'll think American we don't have a strong European ecosystem for artificial intelligence that is why we Host this conference because we want to show that we have talent here we want to bring everything together but currently UI will probably think American or Chinese it won't share European values and you can think about the consequences about this and we want to change this and if you want that your AI your self-driving car is having values from a culture which you have never lived in you don't speak and you don't know but it will act like this why because we train AI so the AI is the replication of the people who trained it and it will be trained in the country W program most often

so this is Rise of AI it's a day to inspire you it's a day to teach you it's a day where you hopefully get enough brain food to discuss to take home I hope that you all have system overload at the end yes I want to give you so much information today with all these speakers um that you have enough to think for a couple of days but rise of AI is also place to meet people to look for investors who want to finance your startup to look for startups you maybe want to finance for potential clients someone to hire someone who did research you want to support you want to work you all here to mingle to network because you're all interested in this topic you all have a connection to this and this is the most important about this so be my last words artificial intelligence is just a tool for us to be more human we humans will stay for a while the super intelligence we are still the ones who do

all the brain work and we just use AI as a tool to extend our will so there is no fear of super AI yet I don't know what will be in 10 to 20 years but the next 10 years I'm pretty sure we will stay the super intelligence in the room thank thank you

Rise of AI 2027 · Berlin · May 2027

Be in the room for the next one.

Ten years of talks started as conversations in one curated room in Berlin. The next edition brings 300 selected leaders together in May 2027.