Davor Orlic | Humane AI

April 17, 2019

Blog

Luc Steels, Icrea Research Professor at the Institute for Evolutionary Biology

Humane AI or Human-Centered AI has many meanings for many people, and it’s a bit unclear at the moment. For me it means that we kind of gain back the let’s say emphasis on mechanisation and on basically treating people like machines. For me Human-Centered AI has to do with meaning and values in our society but also on how we treat other people. I have been very disturbed by the use of AI for manipulation, getting their data and using AI in places where in my opinion it shouldn’t be used. With this project I hope that we can correct this deviation towards using AI to dominate people and to manipulate them, rather than help them, which was the original goal of the development of AI.

My blue sky project would be to bring together the people in Europe who work in arts, design and architecture who actually have a tradition of building human centred technologies or objects. If you look at Italian design of a lamp for example, it looks very simple but actually they put a lot of thought in that and thought about the human use of it and the context of it. I think we need to radically change how we are dealing with AI. I think we can do fantastic things with AI but we have to merge this kind of way of looking at things, the artistic way I would say. To bring that together with the power of technology. At the moment technology is completely in the hands of people that are mostly interested in money and manipulation. We have to change that so that we can benefit from it.

April 16, 2019

Blog

Carles Sierra, Artificial Intelligence Research Institute

More and more we are involved in applications that are involved or have a social side. We have been working on AI for education and healthcare. We strongly believe that artificial intelligence needs to interact with humans, needs to be integrated in mixed societies with humans and other kinds of machines, so the Humane dimension is essential. There cannot be artificial intelligence without a Humane dimension.

My blue sky project is to have people in control. I am not really willing to have let’s say people in the loop, but people in control. That means that we need to integrate a number of technologies that go from normative systems, argumentation and automatic programming with the purpose of achieving what would be my dream, that a community of people interacting through artificial intelligences define what should be their way of interacting and what should be the norms about their interaction, what should be the way artificial intelligence helps them. And through a process of democratic discussion or selection of procedures, those procedures become code, that is the actual code that those artificial intelligence are going to respect in order to satisfy the constraints that humans have in terms of their social interaction, ethics and so on.

April 16, 2019

Blog

Marko Grobelnik, Artificial Intelligence Laboratory, Jozef Stefan Institute

Marko Grobelnik, Artificial Intelligence Laboratory, JSI

The issue of artificial intelligence and humanity can be answered in several ways. One possible view is that Humane AI is kind of a European view to AI versus let’s say American, Chinese, or Russian and so on, which maybe have other focal points, and Humane AI would be closer to a human, so human-centric with ethical values emphasise, but in its core is really has a European dimension in itself. Throughout the discussions that we have this idea is emphasised and I think it’s kind of right. There needs to be a counterpart or counterbalance to these other ways of dealing with AI.

My blue sky project is realistic and its actually happening. In a few short sentences, can we predict future events in society? How can we do it? By understanding society in a causal way. There is this famous statement “history doesn’t repeat itself, but it rhymes”, so can we find the rhymes of society so that we can understand and possibly predict the future of society on a micro level. This would be one of them, but the other area that is in my strong interest is common sense reasoning. Today’s narrow AI which is solving hard problems, but extremely narrow ones, can we broaden this into something which we could technically name common sense reasoning.

April 15, 2019

Blog

Wendy Mackay, INRIA

We are in a situation now where AI is becoming way more powerful, and as it gets more powerful, we need to think about what is the impact on human beings. I am really interested in the problem of how do you combine human beings and artificial intelligence, and the problem that we often have is that we think of AI as a way of replacing human beings and what we really need to do is to think about AI as a way of empowering people. Making human beings better. We don’t want to de-skill human beings but add new creative opportunities for people and AI should be viewed as a n enhancement and not a way of replacing people.

I am really interested in human being and creativity, creative professionals. How do you create an environment where human beings actually control the technology where the system reveals itself to the user in a way that they can then control it and do more things that they could do before? So that is my challenge, how do we use AI in service of human beings and particularly in human creativity.

April 14, 2019

Blog

Catholijn Jonker, TU Delft

AI can do a lot for good and I see a coevolution between humans and technology, and if you think about technology as just being there to make us more comfortable and do all the work for us, then I am afraid that the coevolution of humankind will go into the direction that is basically leading us to extinction. I would like to create artificial intelligence that challenges us, bring us to a higher level of evolution into the direction that we actually want to go. Determining that direction means that we have to talk about it and we have to stay in debate with each other about it but also with the artificial intelligence we are creating. So for the HumaneAI is really about making that combination, how to challenge each other and grow.

My blue sky project is to create methods for mass deliberation. We are all talking about having participants form the community and having all citizens participate, companies participate in our democracy, we also know that direct democracy gives us interesting problems. How can we at least make sure that everyone that is part of our society, which is everyone, can really understand what the debate is about, understand why others see in a different view and we also believe that if your approach is from a value oriented perspective, you see that the others are not demons, they just have understandable value set that maybe you give a different priority to. You might go for privacy, I might go for safety, and because of the difference in value weight, you might make a different decision than I do, but that doesn’t make you a demon, and it doesn’t make me a demon. So how can we work together and understand each other and form that point of view hopefully find new solutions to the problems our society faces.

April 13, 2019

Blog

Felix Schoppa, German Entrepreneurship GmbH

I think the project is really targeting the mind set of current times. AI is coming from everywhere and mostly its rising in China and US and I think we really need in Europe initiatives like this project to have an answer to that and to foster responsible AI, because we do not know if the other parts of the world are going in the same direction as we are.

My blue sky idea would be using AI for VR and AR applications to educate people all over the world to make them empathise in situations that happen in other parts of the world. To have a worldwide education system and to work towards people understanding people in different situations.

April 13, 2019

Blog

Alípio Jorge, University of Porto

Europe has a very important role in pushing the human aspect into artificial intelligence, because we are on one side competing with US and China we have a completely different view of what AI is and different start. In Europa we are by culture working on AI with very strong human drive and this is very important. Not only to be different form the others, but because it’s what we believe in and I think it’s not disadvantage as some people see it. Data protection can be a disadvantage for AI companies on the other hand I think it’s a strong advantage that we can have and that we should have.

Understanding what people do and say and that has to do with language technologies and connecting them with the real world, so finding ways of discovering semantics and actually understanding what text and words say.

April 12, 2019

Blog

James Crowley, Institut Polytechnique de Grenoble

If you look at where money comes from, what drives AI. In America AI is for profit. Google, Amazon, Apple, all these companies are looking to make money. This has led to some perverse effects of the technology taking over from people. If you look in China, right now where the real innovation is, China looks at AI for control, this leads to a ubiquitous vision of what everybody is doing everywhere and an incredible tool for controlling people and it AI for humanity. Our vision is to offer an alternative to AI and AI for money and AI for control. an alternative vision for AI for humanity. We want to make AI that will empower humans, to give them new powers, new abilities, to improve their quality of life and not to replace them.

AI for humanity opens many interesting blue sky projects. My current challenge is to build systems that collaborate and understand people as individuals. That understand what they understand, to see what they see and to see what they don’t see, so that they can offer advice, assistance and services to help them be more effective. These are what we call collaborative intelligent systems.

April 11, 2019

News

HumaneAI project becomes reality

The Humane AI kick-off meeting was held on 11 April 2019 at the CINIQ center in Berlin with all partners attending.

Humane AI, with project number 761758, is funded under the topic “FET FLAGSHIPS – Tackling grand interdisciplinary science and technology challenges” of European Union’s Horizon 2020 research and innovation programme.

The focus is on human-centered AI, with a strong emphasis on ethics, values by design, and appropriate consideration of related legal and social issues. The HumanE AI project will mobilize a research landscape far beyond the direct project funding and create a unique innovation ecosystem that offers substantial return on investment. It will result in significant disruption across its socio-economic impact areas, including Industry 4.0, health & well-being, mobility, education, policy and finance. It will spearhead the efforts required to help Europe achieve a step-change in AI uptake across the economy.

The consortium with 35 partners from 17 countries, including four large industrial members, has defined further steps to implement a full research agenda in Human oriented AI and a strategy to mobilize major scientific, industrial, political and public support for this vision.

The Humane AI kick-off meeting was held on 11 April 2019 at the CINIQ center in Berlin with all partners attending.

April 11, 2019

Blog

Davor Orlic, Knowledge 4 All Foundation

The idea of a human approach to AI seems to be at the center of gravity in Europe. It’s an interesting idea and the room supports it, but there is also this question of who is involved, policy makers, researchers, politicians. We will probably want to look at the choices we have. AI for cyber security, AI for the military versus AI for taxpayers and the general public and the citizens of Europe.

My bluesky project would be to connect all of the machine learning communities across the world into one, a sorts of “machine learners of the world unite” approach. This would mean to build on value but also to protect themselves from other players and paly with other players, such as policy makers, etc. A global south, a global north and a global network.

April 10, 2019

Blog

Michael Klein, VU University Amsterdam

like very much this idea of the Humane AI project because I think technology is never on its own, and especially this is not the case for AI technology, so its always happening in interaction with humans. Within Humane AI we are trying to see how we can study what is happening when people are using AI technology. That’s really relevant. I have heard a really nice summary today when somebody said that the US are doing AI for money, China is doing AI for control and we in Europe are doing AI for humanity. I think that’s very important. We want to achieve something with AI that is good for humanity.

My own research is about understanding human behaviour. There is a lot of knowledge about why people do what they do in psychology, sociology and we also have a lot of technical measures to work on that. I would like to have an integrated framework where we combine all the knowledge and technologies that we have to understand what people are doing and why, that’s my bluesky project for AI.

April 9, 2019

Blog

Virginia Dignum, Umeå University

Humane AI is about people and AI, about developing AI systems in a way that they are fundamentally inserted in our societies in a way that it understands and respects our values and works with us for augmenting capacities to making our life’s better and to make society a better place for all.

My bluesky project in AI is about designing optimal AI, meaning that AI is optimal for people and that is trustworthy, ethical, responsible and will take accountability and transparency into account. The question is how to develop computational tools to implement all these ideas into these AI systems.

Luc Steels, Icrea Research Professor at the Institute for Evolutionary Biology

Carles Sierra, Artificial Intelligence Research Institute

Marko Grobelnik, Artificial Intelligence Laboratory, Jozef Stefan Institute

Wendy Mackay, INRIA

Catholijn Jonker, TU Delft

Felix Schoppa, German Entrepreneurship GmbH

Alípio Jorge, University of Porto

James Crowley, Institut Polytechnique de Grenoble

HumaneAI project becomes reality

Davor Orlic, Knowledge 4 All Foundation

Michael Klein, VU University Amsterdam

Virginia Dignum, Umeå University

Knowledge 4 All Foundation Ltd.

Humane AI on Social Media