Yes, We Care! Results of the Ethics and Natural Language Processing Surveys

Yes, We Care! Results of the Ethics and Natural Language Processing Surveys Karën Fort∗, Alain Couillault† ∗ Université Paris-Sorbonne/EA STIH, † Université de La Rochelle/L3I karen.fort@paris-sorbonne.fr, alain.couillault@univ-lr.fr Abstract We present here the context and results of two surveys (a French one and an international one) concerning Ethics and NLP, which we designed and conducted between June and September 2015. These surveys follow other actions related to raising concern for ethics in our community, including a Journée d’études, a workshop and the Ethics and Big Data Charter. The concern for ethics shows to be quite similar in both surveys, despite a few differences which we present and discuss. The surveys also lead to think there is a growing awareness in the field concerning ethical issues, which translates into a willingness to get involved in ethics-related actions, to debate about the topic and to see ethics be included in major conferences themes. We finally discuss the limits of the surveys and the means of action we consider for the future. The raw data from the two surveys are freely available online. Keywords: Ethics, Moral Buffer, Whistle Blowing Ethics in|by Natural Language Processing Natural Language Processing (NLP), like any other sci- ence, is confronted to ethical issues, both regarding the way science is conducted (plagiarism, reproducibility, trans- parency) and regarding the effects of its results on society. Moreover, the 21st century is witnessing major threats to citizens’ privacy in which NLP techniques play a central role. The time has therefore come to gather and share our reflections on the impact of this evolution on the way we conduct our research. Indeed, NLP applications are often used to analyze and ex- tract information from documents which, by their nature or their content, require thoughtful considerations regarding ethics. This is for example the case for all private content like Emails or suicide letters. It has been taken into account for a long time for medical corpora. Furthermore, the build- ing, transformation or annotation of the (sometimes huge) language resources implies to set up and drive large scale projects involving human resources, thus generating poten- tial issues with regards to working conditions. On the other hand, NLP is called for when it comes to pro- viding tools for de-identifying documents, discovering pla- giarism or performing fact checking, all applications that can be considered as ethically positive. This contrasted image of the domain has shown in the Journée d’études ATALA on ethics and NLP1 and the TALN ETeRNAL workshop2 that were organized last year in France. We decided to run a survey to try to understand better the needs and concerns of the community regarding ethics. A first poll was conducted within the French NLP commu- nity and its success encouraged us to run a second poll at the international level. We first give a rapid overview of the range of actions per- formed in the more or less recent past concerning ethics and 1See: http://www.schplaf.org/kf/JE_ATALA. html. 2See: http://eternal.loria.fr/. NLP. We then detail the results of the …

Content truncated to 500 words