How AI will make us collectively creative

How AI will make us collectively creative

As a managing Director of Idea2Innovation Sweden AB, Jakob is exploring the intersection between human creativity and new technologies and his team. They focus on idea generation and co-creation. In his talk at TEDxÖstersund, he is describing how the new possibilities arising from big data and artificial intelligence will enhance our capacity to collaborate and innovate at scale and make new friends!

We are facing some serious challenges today in the world.. Challenges that need innovations to be solved and we need to do this together. Without creativity, there will be no innovations and without collaboration it’s hard to be creative. It’s really important to face those challenges and collaborate together.

Me and my team has been working with several years creating software to make it easier for people to collaborate and contribute with their knowledge, with their ideas, with their insides and with their contributions.

In our team, we have a computer guy called Hola. Hola is not just a computer guy, actually he is football analyst. He is football fan and has been following the results in the top European League since he started to read books and papers.

Since he is the computer guy, he dreamt about analyzing the data and being able to beat bookmakers. Isn’t that the ultimate dream to make a computer program that actually could beat bookmakers?

He started on with this hobby project three years ago. He started gathering as much data as it possibly could when it comes to football games in the European Premier League’s. He downloaded the data from 30,000 football games all over Europe and by speaking about data it is obviously which team did meet, who won, how many scores, who did play whom and who was away, who made this course, when they made this course, how many people were there on the field, (11 of course) but they change and all those what was a better like, how many people further in the arena, did some of the teams have disadvantages etc. all this data it downloaded on his computer. To analyze all this data, he used a kind of Artificial Intelligence called Deep Learning.

Deep Learning is a very specific part of Artificial Intelligence where the mathematical algorithms that are solving problems. Mathematical algorithms are solving problems more or less in the same way as our brain.

If we think about our own brain which is maybe a little bit hard to do but if we try to imagine how we solve problems. A problem could be for example to see something that we have never seen before which we don’t actually know what it is. To solve that problem, our brain is cooperating inside between neurons.

We have trillions of neurons in our brains and every neuron is trying to take a little bit of this problem and solve it. Our brain is actually cascading the problem and solving it piece by piece putting those pieces together to something that we can understand. That’s how the brain is working trillions of neurons.

Deep learning is a way to do it mathematically according to the same principles. One of the best examples is Google brain. In 2012, they experimented with building large network of processors connected together. There’s sixteen thousand processors connected together and more than 1 million connections between them and they downloaded 10 millions of pictures from YouTube videos and without instructions they told the computer “find something” and without any supervision no human being ever actually interacting the computer program. It took the computer program three days to discover the concept of a cat. It actually can find cats in those movies and could also tell the programmers “Look I have found a cat it looks like this”. This was real evidence that computers actually can learn.

This technique was used for trying to predict the outcome of football games. At the same time, we were working with ideas in our team. Our computer program makes it possible to find ideas from the crowd so called “crowdsourcing”.

We had a lot of discussions about how to find the really good ideas if you’re sitting with a bunch of thousand or ten thousand ideas, how can you find the good ones. We made research together and at one meeting over, I said “you know my hobby being a football analyst, I have kind of succeeded at least in any way and maybe we can use this technology to find good ideas”.

We started a project. Well, we try to see if it is possible to find good ideas using deep learning algorithms and basically what the algorithms did was that it analyzed the idea as such the content of the ideas, the title of the ideas, the description of the ideas, what people thought about ideas in the comments field if they’re like ideas and We also analyze the context of the ideas who came with the idea, who have seen idea, and who have interacted idea. We tried to make a computer program that could predict which ideas were likely to be very very good.

You can wonder what happened with our algorithm could be actually detect really good ideas? The answer is sort of yes. We could quite easily detect ideas that ought to be very good but you must need to look at them but we did not find a way to find disruptive ideas and basically the same reason for why algorithm doesn’t beat the bookmakers and why we can’t find disruptive ideas.

That’s because it’s impossible to predict the unpredictable. So we have learned that we can make generalizations but we cannot predict something that is unpredictable. So finding disruptive ideas real artificial intelligence is at least yet not done..

But also He said another thing, he said to me do you know where my algorithm actually did the giant leap to being quite good and I said no I don’t it was when I added twitter feed to the algorithm.

So what will indeed also defeated Twitter into the algorithm and Twitter from fans twitter from coaches twitter from analysts into the algorithm and that made us think that when we are adding the human touch to the data we have passive data but we can add human data then we get something that mathematics don’t can do by yourself because human data input from humans;contains, creativity, imagination, and all this kind of human stuff that computers can’t understand by themselves. so where are we heading? I believe that there is something in social data that will really enhance our ability to cooperate so give me just of some seconds talk about social 1994 internet as we know it came with a netscape navigator that made it possible for us to actually browse the web. It took 10 years to reach 1 billion connected users. Today we have 3.5 billion users connected to the internet. It’s forty percent of the world’s population. In the developing countries that are coming now really really fast more than ninety percent of internet access is from mobile phones. You are in a tremendous speed two order really connected community globally.

One of the driving forces is of course social technologies and when you look at the figures when it comes to social technology it’s really astonishing. it’s more than one-and-a-half billion people active on Facebook every month and in China they have something called the WeChat they don’t use facebook there’s more than 1 billion people on WeChat. every minute there are 400 hours of videos uploaded on YouTube. since I started this speech there’s videos that takes a human being 20 weeks to look at since I started this speech 10 minutes ago. so you really need the computers to browse all this data that is produced in social technologies there is two and a half million likes on Instagram every minute. And we cooperate and collaborate around every topic you can find about on Pinterest fifty percent of all the interactions is food-related. It’s a variety of their of the things that we discuss on the net.

What could be really interesting here that our behavior on social technologies is actually I cannot say recorded but it’s we collect all the data and of course there is an integrity aspect of this. It’s really important for us to take into consideration how it looked on privacy and integrity. But apart from that all this data means that we can actually make a map about people’s relations. Facebook released in 2007 something they call the social graph and the social graph is basically a map about who knows who, who interacts with who, and who is interested in what and this social graph is accessible for every computer programmers who want to make programs and interact with people through the social graph API and a lot of a large social networks they have this kind of social graph technology.

So putting those together we can analyze ideas and really see what is a content of this idea, what is good, and what is missing, and who should take a look on this idea, to take it further on and we also have the knowledge about the people who knows what, who is interested in what, who could be interested in contributing to this idea.

This is where everything will be together so what will happen is that we will automatically have a kind of facilitation or idea development in our networks. Inside our companies in the civil organizations and in on the public internet we will have automatic facilitation of discussions and idea development and try to find those new ideas where really can make through and I think that’s a very very big thought that everyone on the planet that could help me to contribute to my idea. Could I reach with this kind of working so this is how artificial intelligence will help us to matchmake each other.

I think that’s more important than ever. If you listen to an astronaut who has been out in the space, a lot of them say two things that when I was a kid all the globe’s had those kind of borders but when I look at the globe from the space I see no borders it’s obvious for me that we live on the same planet together and they also say another thing. That earth looks fragile if you look at the surface of the earth the tiny the tiny space and then it’s the ultimate everything it’s serious challenges we have poverty, water supply, food supply, global warming, civil rights.

We really need to find break through innovations to solve those challenges and I believe that the use of social technologies and artificial intelligence is one approach that will help us because it will be possible for us to matchmake people with ideas and knowledge and resources to solve those problems together and create new innovations.


Social Media Share

Leave a Reply

Your email address will not be published.