by Sabine Theresia Köszegi
On 25 April 2018, the Europaen Commission published a Communication in which it announced an ambitious European Strategy for Artificial Intelligence (AI). The major advances in AI over the last decade revealed its capacity as a general-purpose technology and pushed inventions in areas of mobility, healthcare, home & service robotics, education and cyber security, to name just a few. These AI-enabled developments have the capability to generate tremendous benefits not only for individuals but also for the society as a whole. AI has also promising capabilities when it comes to address and resolve the grand challenges, such as climate change or global health and wellbeing, as expressed in the United Nations Sustainable Development goals. In competition with other key players, like the United States and China, Europe needs to leverage its current strengths, foster the enablers for innovation and technology uptake and find its unique selling proposition in AI to ensure a competitive advantage and a prosperous economic development in its Member States. At the same time, AI comes with risks and challenges associated to fundamental human rights and ethics. Europe therefore must ensure to craft a strategy that maximizes the benefits of AI while minimizing its risks.
The Commission has set out an interwoven strategy process between the development of a European AI Strategy and the development of a Coordinated Action Plan of Member States (hosted under the Digitising European Industry framework). The publication of the European policy and investment strategy on AI is envisaged for Summer 2019. To support this strategy development process and its implementation, the Commission has called for experts to establish a High-Level Expert Group on Artificial Intelligence (AI HLEG). Following an open selection process by DG Connect in spring 2018, the Commission has appointed 52 experts encompassing representatives from different disciplines of academia, including science and engineering disciplines and humanities alike, as well as representatives from industry and civil society. As an expert in labor science and with a research background in decision support systems, I was selected to join the exciting endeavor to lay the foundations for a human-centric, trustworthy AI in Europe that strengthens European competitiveness and addresses a citizen perspective to build an inclusive society.
Our mandate includes the elaboration of recommendations on the policy and investment strategy on ethical, legal and societal issues related to AI, including socio-economic challenges. Additionally, we serve as a steering group for the European AI Alliance to facilitate the Commission’s outreach to the European society by engaging with multiple stakeholders, sharing information and gathering valuable stakeholder input to be reflected in our recommendations and work.
On 18 December 2018, we proposed a first draft on “Ethics Guidelines towards Trustworthy AI” to the Commission, setting out the fundamental rights, principles and values that AI has to comply with in order to ensure its ethical purpose. Additionally, we have listed and operationalized requirements for trustworthy AI as well as provided possible technical and non-technical implementation methods that should provide guidance on the realization of trustworthy AI. This draft on ethics guidelines is currently in a public consultation process in the European AI Alliance platform. Through this engagement with a broad and open multi-stakeholder & citizen forum across Europe and beyond, we aim to secure the open and inclusive discussion of all aspects of AI development and its impact on society. The finalised draft will be formally presented in the First Annual Assembly of the European AI Alliance in Spring 2019.
To advise the Commission with regards to the European policy and investment strategy, we are currently preparing a set of recommendations on how to create a valuable eco-system for AI in Europe in order to strengthen Europe’s competitiveness. The draft document of recommendations should be published in April 2019 and will undergo a public consultation process as well. The recommendations will primarily address European policy makers and regulators but also relevant stakeholders in Member States encompassing investors, researchers, public services and institutions. I would like to use the opportunity, to invite the readers of ERCIM News to engage in the European AI Alliance (see the link below) and to contribute your expertise and input to our policy and investment recommendations.
The complexity of AI-related challenges requires to set up a problem-solving process with highest information processing capacities that allows to consider different perspectives and to resolve conflicts of interest between different stakeholders. It can easily be imagined that our discussions as an inter-disciplinary expert and multi-stakeholder group are intense, difficult and at times emotional. In difficult situations, I remind myself of our commitment to the following statement in our ethics guidelines: “Trustworthy AI will be our north star, since human beings will only be able to confidently and fully reap the benefits of AI if they can trust the technology.”
Sabine Theresia Köszegi is Professor of Labor Science and Organization Institute of Management Science, TU Wien, Chair of the Austrian Council on Robotics and Artificial Intelligence, BMVIT, and Member of the High-Level Expert Group on Artificial Intelligence of the European Commission.