Blog Posts

Viviana Gropengiesser, German Entrepreneurship GmbH

Viviana Gropengiesser, German Entrepreneurship GmbH

I think an important part of what we are trying to do at the Humane AI project is that we need to bring together all different kinds of entities to understand a lot of different aspects of a very complex topic and increase the involvement of very different players and stakeholders in the field, to create something that the entire society and the industry, economy and science can actually support and stand behind

If I had unlimited resources my dream would be to really able to educate on all levels what opportunities we have and what is happening and bring people together to create something that everyone participated in or can participate in the future and change the future to something we all want to live in.

Albrecht Schmidt, LMU Munich

Albrecht Schmidt, LMU Munich

For me the Humane AI project is really bringing together people from interaction, artificial intelligence, machine learning, machine vision, human computer interaction to in order to really create something new. How can we use artificial intelligence for the human, how can we make it humane, how can we create new things that take artificial intelligence and that are valuable for humans?

It’s really exiting form me to be in the field, and one of the things we are working on is how to amplify human ability, how to amplify the human mind and perception through artificial intelligence technologies. How do we closely integrate our abilities as humans to perceive, think and create with the technology and use AI as a tool to make us smarter, to perform things we couldn’t do before.

Helena Lindgren, Umeå University

Helena Lindgren, Umeå University

Humane AI is extending AI as technology to produce intelligent systems also to apply it to situations where people live in work. It has to do with people, environments and where we interact and work with AI in society.

For Europe, for people, for society it would be along with what interests me, cognitive tools for individuals is a system that can help individuals cope with everyday  living and working, managing stress and everyday conflicts and interests in your daily life.

Fosca Giannotti, Research ISTI-CNR

Fosca Giannotti, Research ISTI-CNR

AI has transformed and will transform even more society and the way people live. The way we buy, the way we do social relationship over the Internet, everything is very much now mediated by smart technology that learns from the usage of other customers how to make you more reactive in terms of likes and business. This kind of technology that facilitates our lives in a sense, is also creating bubbles. Its forcing us, to talk to people that are too similar to us, buying things that are very close to us. It’s changing society. And sometimes the way these changes happen is that the way the algorithms are learning and the data that we leave on the Internet. The large success of current AI is mostly because of the enormous availability of data, which are traces very often left by humans.  Data contains all the issues that society has, including the bad parts. Our algorithms might learn by data that contains bias. Sometimes those bias turn out to be discriminations on vulnerable people or learn language that might not be politically correct. We need an AI which is helping society also growing in terms of education and civil interactions.

My bluesky project is very much related to my bigdata vision of AI. It’s about having an AI and machine learning methods that always offers a human understandable explanation of the decisions they suggest, by taking the human in the loop and express such explanation in language that different kind of users may have depending by their context.

John Shawe-Taylor, University College London, UNESCO Chair in AI

John Shawe-Taylor, University College London, UNESCO Chair in AI

John Shawe-Taylor, University College London, UNESCO Chair in AIThe jump into AI has to some extent jumped into a technology that can replace humans, a technology that can influence humans and to some extent undermine humans. So there has been obviously many important advances, and very exciting development in the technology, but it hasn’t always been thought through what the effect of this technology on humanity will be and to ensure that the potential negative effects are addressed or avoided. This is completely understandable as many of the negative effects were unforeseen.

The ability to influence people and the way we’ve seen this possibly misused in election processes is something that I think people would not have foreseen and for that reason there is no guilt on the part of companies for failing to address these issues upfront, perhaps they have been slow to respond in many cases, but that really comes down to their business model. A business is in many instances responsible in many respects for ensuring profit for their shareholders.

Of course it should take social responsibility seriously and perhaps that’s being sometimes taken on board a little later than it would be desirable. However, I think this HumaneAI project is to address these issues head on and is to take those questions through to the fundamental research that is required to address them to understand them properly and to think about their implications for humanity as a whole. And there I think we are potentially hitting on incredibly interesting questions.

Questions about “what it means to be a human being”, “what is now the essence of being human”, perhaps it isn’t just intelligence. If we can reproduce that intelligence in a machine, then why are we different from a machine? We have to ask that question, and I think that is one of the questions that Humane AI can shed light on by understanding the implications of the way humans think and do things, and experience and the way AI system do things and don’t experience, for example.

One of the beauties of Europe is its cultural diversity, and to some extent I am a great believer in not undermining that diversity by unification. I think we want to encourage that diversity, individuality in communities, obviously not at the expense of unity in the sense of understanding each other. But that’s again at the heart of hitting at one of the problems that AI has created, which is this potential for information bubbles, where people only ear the information the news the perspective that agrees with their current thinking. And that is not the kind of individuality that we want to grow in Europe.

We want to grow an individuality that respects differences, understands differences, celebrates differences, but equally celebrates individuality. So we don’t get rid of differences by creating a uniform mish-mash, we create a collaborative set of experience that is bringing together very different communities but ones that respect and encourage that individuality in others. That requires tools that will enable people to communicate across cultural divides and to understand different cultures and see what the other side thinks and see what is the difference. We can have different communities appreciated and supported in their individuality.

Chiara Boldrini, National Research Council of Italy

Chiara Boldrini, National Research Council of Italy

Humane AI is a very important project for all European countries, because we all know that in the US, artificial intelligence is mostly business oriented, you have to make money out of it. Other countries are more focused on how to control people, in order to know what they are doing and control their actions. In Europe according to the European spirit we are focused don what AI can do for people, how we can improve their way of life and also how can we make sure that AI is not detrimental to peoples lives, but to improve the quality of life, jobs and give them possibilities that they currently don’t have.

A bluesky research project that would fill a huge gap that there is currently in artificial intelligence but also in general in data mining is actually the lack of data. We talk a lot about bigdata, but there are not many data or its proprietary and it’s not very easy to study and research on this topic. For example, if you study social media, a lot of this data is owned by companies, so as a researcher you can only observe a small subset in that world.

Paolo Traverso, Fondazione Bruno Kessler

Paolo Traverso, Fondazione Bruno Kessler

Humane AI is the distinguished vision of Europe for AI. When we say something like this in the US they make AI for big business, in China they make it to govern. We make AI, we want to do research in AI for the people. Human AI is to improve the life of the people. I think that to do this you need a distinguished and different technology and research. This is something that can be done only by Europe.

At the moment there are two main trends in AI, one is more from learning from data, the other is more from on modelling the world, systems and machines. I believe the future will be an integration leading to integrative AI. Both components will be put together. We need people with different competencies in different branches of AI to work together to do something better and stronger than before.

Aart van Halteren, Philips, Vrije Universiteit Amsterdam

Aart van Halteren, Philips, Vrije Universiteit Amsterdam

As a company it’s really important that we do AI research, that its really done in a responsible way, so that we don’t end up in all kinds of discussions that would put a company in a bad situation. I think it’s really important that we get a group of people together that work with that type of objective.

My bluesky project would be to build this cognitive exoskeleton that would allow you to as a patient to really manage your health in a much better way than you would be able to do without it. Staying with your own personal values and be able to get the right support when you need.

Raja Chatila, Sorbonne University

Raja Chatila, Sorbonne University

Humane AI is a project or an endeavour shaping the whole AI research of the future. We want Human-Centered AI, what does this mean? It means that when we develop this technology, an over reaching technology and very important for our future, we should really keep human in the centre which means this technology is there to serve humans, to benefit humanity, individuals and society. It’s not there to be developed by itself. It’s there to increase our wellbeing and our understanding of what intelligence is. It’s really there to help us to grasp better our future. It’s a formidable tool that should be developed correctly to help us increase our benefits.

Every researcher in AI wants to understand what is intelligence. We have different approaches to that. I think to understand intelligence, we need to understand how intelligence interacts with the real world, because the real world puts limits and challenges. The world is changing its complex and very difficult to understand, and this is what intelligence is about, making sense about the real world. It’s not working in an abstract and simulated world. It really has to address both the real time constraints and the evolving complexity of the world. My take on what would be a real blue sky project in AI is to make it interact with the really world and understand how it can do that efficiently.

Philipp Slusallek, German Research Center for Artificial Intelligence

Philipp Slusallek, German Research Center for Artificial Intelligence

We are in this situation where AI has been a very fragmented area in Europe, which is very different, for example Google brings things together, Baidu and the Chinese government on the other side while Europe has a lot of groups all over the place with perhaps an exception at DFKI. We felt that if we really want to move things and really be competitive we need to bring the community together. Even the biggest organisations like DFKI which is the largest research centre in AI worldwide, we can’t do this alone.

This is why we started CLAIRE, which is a confederation of labs in AI in Europe. We have 260 or more labs, almost 3000 individual supporters., so there is a really momentum building and under this big umbrella we are arguing for this network of labs and a CERN for AI as a trademark. In that context of Claire one big element is research and this is why we started this Humane AI as the really big research initiate in the Claire context in Europe tor really create major breakthroughs with all the top people in AI in Europe and make sure they don’t go to the US and maybe even bring some of them back to Europe.

My specific project is the digital reality, it’s the idea of modelling the world. We do this mostly for autonomous driving mostly right now. We are modelling how pedestrians are moving. As humans we never this if a pedestrian will move or stop at the edge, we know, right, we see the subtle changes. Can we model that to create artificial synthetic scenes that incorporate these fine subtle changes in human motion so we can train them into AI systems. We can use these models to create test cases to make sure that a future autonomous car is actually detecting this and will not kill a pedestrian. Its really the validation of AI systems through modelling the environment, sing this for training and other purposes. It’s a large scale thing and applies to any AI system where you can model the environment it interacts, but autonomous driving is really a key element that we are looking at right now.

Samuel Kaski, Finnish Centre for Artificial Intelligence

Samuel Kaski, Finnish Centre for Artificial Intelligence

It’s important that there is a third route of AI which is not purely big company interest driven or government driven in the sense of China. There needs to be a third route which is the European way of talking the humans seriously. Technically, in the sense that AI is able to work with real people, but also in the sense that the AI is humane, and takes into account the human interest.

I would want to have a machine, an AI that is able to assist humans in design. By design I mean as in artistic design but also designing of engineering systems, traffic systems, designing which drugs to give people based on what has been measured of the people, etc.

Andrzej Nowak, University of Warsaw

Andrzej Nowak, University of Warsaw

AI is most likely or arguably the most important discovery in science and change in technology. On the one hand it represents a real threat to humanity, if it goes the wrong way, on the other it is perhaps one of the best things that ever happened if it takes the right root. I think the dangerous root is that it can be a competition to beat human intelligence and replace it and take away jobs and limit our freedom. On the other hand, is one of the most powerful tools that can enhance our knowledge, increase our freedom and understand our capacity to understand the world around us, our knowledge and our freedom. Humane AI is trying to develop AI in a way that it starts working for humans, that is enhances human capacity rather than takes limits them and takes them away.

One thing that is happening with the development of the global interconnected world and influences over the Internet and social media, is that our freedom of choice is being limited. We are pushed by advertised products, we are supposed to make choices that are manipulated, etc. I think my bluesky project is to make this equation more even and to think about how AI might be used to increase our possibilities of choice and how to increase our access to unbiased in formation and balance this information mess we are living in. In a way with the general vision I have for AI, which I is to make humans better understand their environments, more capable of free choice, and more aware of their consequences.