More than words, more than avatars. Introducing the DeafSafeAI ethical principles

An increasing number of research projects aim to develop translation interfaces between spoken languages and sign languages and/or focus on sign language technologies. In line with a form of technological solutionism (Morozov 2014), amplified by media acclaiming breakthroughs in communication for Deaf people, we see a proliferation of apps, products, and prototypes of great interest that rarely involve Deaf perspectives. This shifts agency away from Deaf individuals, placing it instead in the hands of tech companies and research institutions (Tucker, 2017). The lack of involvement, not only in workgroups but also in actual decision-making processes, introduces bias in both research and the practical application of these interfaces (De Meulder 2021; Desai et al. 2024; Fox et al. 2023). While artificial intelligence (AI) raises global discussions and leads to the production of rules and regulations, such as the European AI Act 2024/1689 and the 2023 U.S. President’s Executive Order on the Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence, little attention has been given to sociotechnical systems, particularly as they pertain to translation, sign language processing, and sign languages. While AI holds valuable potential, uninformed and unethical use contribute to increased discrimination along with many ethical issues, risks, and opportunities for Deaf individuals. To bring these issues to the forefront, the Coalition on Sign Language Equity in Technology (Co-SET) was established with the goal to ensure that Deaf perspectives on the use of automated or automatic interpreting via AI are thoroughly integrated into policy recommendations. This presentation will introduce the concept of #DeafSafeAI, the 4 Principles for SAFE AIxAI developed by Co-SET, and the fundamental points that, in our view, should guide discussions on the ethical and fair development of sociotechnical systems. References de Meulder, M. (2021). Is “good enough” good enough? Ethical and responsible development of sign language technologies, in «Proceedings of the 1st International Workshop on Automatic Translation for Signed and Spoken Languages (AT4SSL) », 12–22. Virtual: Association for Machine Translation in the Americas. Deaf-Safe AI. A legal Foundation for Ubiquitous Automatic Interpreting (2024). Desai, A., de Meulder, M., Hochgesang, J.A., Kocab, A., Lu., A.X. (2024). Systemic Biases in Sign Language AI Research: A Deaf-Led Call to Reevaluate Research Agendas, in «arXiv:2403.02563». Fox N., Woll, B., Cormier, K. (2023). Best practices for sign language technology research, in «Universal Access in the Information Society», September. Morozov, E. (2014). To save everything, click here: the folly of technological solutionism, New York, PublicAffairs. Tucker, B. (2017). Technocapitalist Disability Rhetoric: When Technology is Confused with Social Justice, in «Enculturation: A Journal of Rhetoric, Writing, and Culture»