Periodic Reporting for period 1 - Humane AI (Toward AI Systems That Augment and Empower Humans by Understanding Us, our Society and the World Around Us)
Reporting period: 2019-03-01 to 2020-04-30
Throughout Europe, there is a current consensus that we need to establish a European brand of AI that is human-centric, with trust, ethics, European values and to the benefit of European citizens and the society at its focus. There is also consensus that this cannot be achieved through a regulation that hinders innovation, but must be instead a motor of innovation driving European researchers to develop new unique solutions—which therefore will also have a unique potential to spin-out innovation from the project and thus generating an increased economic activity A key challenge is that such solutions cannot be found by working within the traditional AI silos, but instead require breakthroughs at the interfaces of various areas of AI, HCI, cognitive science, social science, complex systems, etc.
Artificial intelligence (AI) will bring forth changes that will be much more profound than any other technological revolution in human history. Depending on the course that this revolution takes, AI will either empower our ability to make more informed choices or reduce human autonomy; expand the human experience or replace it; create new forms of human activity or make existing jobs redundant; help distribute well-being for many or increase the concentration of power and wealth in the hands of a few; expand democracy in our societies or put it in danger. Europe carries the responsibility of shaping the AI revolution.
The main objective of the preparatory action was to bring together a unique community which has the expertise both within these silos and at the interfaces between them, and can address those challenges. Furthermore, the aim was to develop a research agenda and a long term strategy to address the key gaps in knowledge that prevent the political vision of a human-centric, European brand of AI from becoming reality.
Artificial intelligence (AI) will bring forth changes that will be much more profound than any other technological revolution in human history. Depending on the course that this revolution takes, AI will either empower our ability to make more informed choices or reduce human autonomy; expand the human experience or replace it; create new forms of human activity or make existing jobs redundant; help distribute well-being for many or increase the concentration of power and wealth in the hands of a few; expand democracy in our societies or put it in danger. Europe carries the responsibility of shaping the AI revolution.
The main objective of the preparatory action was to bring together a unique community which has the expertise both within these silos and at the interfaces between them, and can address those challenges. Furthermore, the aim was to develop a research agenda and a long term strategy to address the key gaps in knowledge that prevent the political vision of a human-centric, European brand of AI from becoming reality.
A key concern in addressing the main objective has been the re-alignment of the project form the original vision of a preparatory action for a 10 year, up to 1-Billion Euro Flagship proposal to a more broadly and vaguely defined project aimed at building and sustaining a community and providing input for the Horizon Europe program.
When the HumanE AI proposal was written and submitted, it was to be one of six projects that were to prepare competing proposals from which 2 Flagships would be selected. Already during the GA signing process it became clear that the Flagship program will not continue in its original form, although the possibility of some sort of smaller funding scheme, for which the 6 projects could compete was left open. However, three weeks after the start of the project during a meeting with all project coordinators the 6 preparatory actions were formally informed that there will be no funding to compete for and thus no proposal to prepare. Instead, the projects should focus on community building, informing and influencing the Horizon 2020 definition, building public awareness and support for the general agenda and looking for their ways to maintain and finance the initiatives beyond the one-year duration of the preparatory action.
Clearly, even without a Flagship as a long-term prospect, the project has been a tremendous opportunity to build a community to advance the vision of Human Centric AI on a European level. As summarized in particular in D7.4 (Results Highlights) the consortium has taken this opportunity to the fullest extent. Nonetheless the re-targeting of the project with the final direction known only nearly a month (March 25th 2019) after the start of a project originally lasting 12 months from March 1st 2019 has been a significant challenge. In total, we had 71 PMs distributed over 50 groups for 18 deliverables to be completed within 1 year. This was a conscious design of the original proposal. The assumption was that the project funding is mostly for travel, coordination, and the actual writing etc., whereas most of the conceptual work will be done as the partners’ contribution to the unique opportunity of being part of a Flagship proposal.
As the project was re-aligned without the prospect of a Flagship, or any other largescale funding to directly compete for several problems arose:
(1) Many deliverables and objectives that were perfectly conceived as building blocks of a Flagship in many cases had a much less clear role within a broader scope of a project directed at general agenda building, input to the Horizon Europe program and community building.
(2) A key task unplanned for in the original proposal became the re-definition of the aims including in particular in particular looking for appropriate funding possibilities to maintain the community after the end of the preparatory action. Most of the other work such as the definition of the research agenda, impact strategies etc. were dependent on this re-definition. Given that the final change in the scope was communicated to the consortium nearly a month after project start, this has been a source of understandable delays.
(3) The expectations of partners investing significant effort of senior personnel beyond project funding was a much more difficult proposition after the prospect of a Flagship proposal was taken away and partners had to re-direct their resources towards other funding opportunities.
Still, the project has been very successful establishing the HumanE AI movement as a key European community with respect to developing and implementing the Commission’s vision of a Human Centric, ethical brand of European AI. By laying the groundwork for the successful HumanE AI net proposal, we have made sure that the ideas and the community will continue to exist and have impact beyond the 1-year duration of the preparatory action. We have also been able to bring together a unique community from different areas of AI, HCI, and social sciences. HumanE AI has also made a big contribution to overcoming the fragmentation within the AI community.
When the HumanE AI proposal was written and submitted, it was to be one of six projects that were to prepare competing proposals from which 2 Flagships would be selected. Already during the GA signing process it became clear that the Flagship program will not continue in its original form, although the possibility of some sort of smaller funding scheme, for which the 6 projects could compete was left open. However, three weeks after the start of the project during a meeting with all project coordinators the 6 preparatory actions were formally informed that there will be no funding to compete for and thus no proposal to prepare. Instead, the projects should focus on community building, informing and influencing the Horizon 2020 definition, building public awareness and support for the general agenda and looking for their ways to maintain and finance the initiatives beyond the one-year duration of the preparatory action.
Clearly, even without a Flagship as a long-term prospect, the project has been a tremendous opportunity to build a community to advance the vision of Human Centric AI on a European level. As summarized in particular in D7.4 (Results Highlights) the consortium has taken this opportunity to the fullest extent. Nonetheless the re-targeting of the project with the final direction known only nearly a month (March 25th 2019) after the start of a project originally lasting 12 months from March 1st 2019 has been a significant challenge. In total, we had 71 PMs distributed over 50 groups for 18 deliverables to be completed within 1 year. This was a conscious design of the original proposal. The assumption was that the project funding is mostly for travel, coordination, and the actual writing etc., whereas most of the conceptual work will be done as the partners’ contribution to the unique opportunity of being part of a Flagship proposal.
As the project was re-aligned without the prospect of a Flagship, or any other largescale funding to directly compete for several problems arose:
(1) Many deliverables and objectives that were perfectly conceived as building blocks of a Flagship in many cases had a much less clear role within a broader scope of a project directed at general agenda building, input to the Horizon Europe program and community building.
(2) A key task unplanned for in the original proposal became the re-definition of the aims including in particular in particular looking for appropriate funding possibilities to maintain the community after the end of the preparatory action. Most of the other work such as the definition of the research agenda, impact strategies etc. were dependent on this re-definition. Given that the final change in the scope was communicated to the consortium nearly a month after project start, this has been a source of understandable delays.
(3) The expectations of partners investing significant effort of senior personnel beyond project funding was a much more difficult proposition after the prospect of a Flagship proposal was taken away and partners had to re-direct their resources towards other funding opportunities.
Still, the project has been very successful establishing the HumanE AI movement as a key European community with respect to developing and implementing the Commission’s vision of a Human Centric, ethical brand of European AI. By laying the groundwork for the successful HumanE AI net proposal, we have made sure that the ideas and the community will continue to exist and have impact beyond the 1-year duration of the preparatory action. We have also been able to bring together a unique community from different areas of AI, HCI, and social sciences. HumanE AI has also made a big contribution to overcoming the fragmentation within the AI community.
The project has managed to re-align itself with the new direction after the Flagship cancelation, and consolidated a strong interdisciplinary community focused on making the Commission’s vision of a unique European brand of Human-Centric, ethical, trustworthy AI that will enhance human capabilities and empower European citizens a reality. In doing so, we have reached out beyond the scientific community to embrace European industrial champions in key strategic areas: from Industry 4.0 through Health to FinTec. We have developed the corresponding research agenda and disseminated it widely both within the scientific community and to the relevant policy- and decision-makers. We have developed a strategy for maximizing the socio-economic impact and creating innovation. Finally, we have come up with new strategies for maximizing synergies in large consortia with limited resources through micro-projects. All that has led to a successful proposal for the HumanE AI Net follow-up project within the ICT48 call that will carry on the Humane AI legacy and ensure that the project will have a long-lasting impact beyond the current one-year “preparatory” action.