by Claude Kirchner (Inria) and James Larrus (EPFL)
Science is in revolution. The formidable scientific and technological developments of the last century have dramatically transformed the way in which we conduct scientific research. The knowledge and applications that science produces has profound consequences on our society, both at the global level (for example, climate change) and the individual level (for example, impact of mobile devices on our daily lives). These developments also have a profound impact on the way scientists are working today and will work in the future. In particular, informatics and mathematics have changed the way we deal with data, simulations, models and digital twins, publications, and importantly, also with ethics.
Ethics in research has been a field of inquiry since antiquity, in particular in health science and more recently in physics. Today all scientific disciplines, from medicine to biology, humanities, informatics, mathematics, physics and chemistry are troubled by ethical issues. They have been particularly popularized by dilemmas posed by the rise of machine learning, AI, and autonomous mechanized entities such as drones, robots, and vehicles.
In the fields of informatics and mathematics, there is a surge of interest in understanding and studying emerging ethical issues and methods for teaching them. Scientists are currently engaged in dialogues about these ethical issues with peers as well as the general population. Professional organisations, such as ERCIM, Informatics Europe, ACM, IEEE and IFIP, are discussing these issues and many advisory documents are being produced at the national and international levels. Let us point to a few examples of such contributions [3, 2, 1].
Of course, ethics in research is fundamental, but it is only one of the corner stones for the investigation of the ethical consequences of the impact of sciences, technologies, usages, and innovations on the digital evolutions. Ethical issues should be investigated everywhere in the world and in particular in Europe to allow the appropriation by everyone, person, organization, and company of the digital advances and transformations that are arising. We are indeed in a situation similar to bioethics in the 1980’s with the necessity to set up ethical committees to consider the consequences of the digital innovations and transformations. Initiatives like the Montreal Declaration [L1], addressing “the responsible development of artificial intelligence, whether it is to contribute scientifically or technologically, to develop social projects, to elaborate rules (regulations, codes) that apply to it, to be able to contest bad or unwise approaches, or to be able to alert public opinion when necessary” are gaining visibility and, we hope, the necessary impact.
To contribute to raising awareness, this section, jointly coordinated by ERCIM and Informatics Europe, features four contributions on “ethics in research”, focusing on informatics and mathematics and the possible interactions with other scientific disciplines. The contributions develop the following topics:
- how informatics, logic and mathematics can help to understand ethics in machine learning techniques;
- why and how reproducible research should be central to research developments;
- education in ethics and scientific integrity is now recognized as a priority, but how should we organize the training for doctoral candidates;
- how shall we organize research to avoid waste and bias in the exploration of scientific knowledge?
These contributions address a small but important part of the overall ethical issues, and we believe it is important and urgent to set up a joint working group between ERCIM and Informatics Europe, to contribute to the development of ethics research in our fields. We invite scientists to join the initiative by contacting us.
Link:
[L1] https://www.montrealdeclaration-responsibleai.com
References:
[1] AIHLEG. Draft ethics guidelines for trustworthy AI [online]. December 2018. https://kwz.me/hdq
[2] CERNA. Research Ethics in Machine Learning. Research report, February 2018. https://hal.archives-ouvertes.fr/hal-01724307.
[3] J. Larus, et al.: “When computers decide: European recommendations on machine-learned automated decision making”. Technical report, New York, NY, USA, 2018. https://dl.acm.org/citation.cfm?id=3185595/
Please contact:
Claude Kirchner, Inria, France
James Larus, EPFL, Switzerland
