The current discourse on AI and ethics is both broad and rich, but there are some structural and fundamental limitations. A recent paper published in Scientific Reports suggest that artificial intelligence is better described as an ecosystem of socio-technical systems where some of the characteristics we should ask of a responsible ecosystem of intelligent systems include being sensitive to ethical and social questions, and promoting the use of technology that support the wellbeing of both humans and our environment.

The current AI discourse provides detailed insights into many of the ethical issues and concerns that AI technologies can raise, and the mechanisms we can use to address them. But how can we assess the impact that this discourse has? Bernd Stahl, Ethics Director of the Human Brain Project and Professor of Critical Research in Technology at Nottingham University suggests one way forward; using the concept of responsibility. According to him, at the moment the discussion about AI is very much focused on identifying specific relationships for responsibility, based on very specific issues. And although those kinds of more individual responsibilities are indeed important, the approach becomes limited as AI is not one distinct and technology with distinct boundaries, but perhaps more accurately described as an ecosystem of socio-technical systems. If we conceptualise AI this way, then the question changes to what kinds of characteristics we would like these system(s) to have.

“Ethicists, scientists and technical experts are very aware of the ethical and social challenges that these technologies raise. Using this line of argument would make the theoretical discussion of ethics and AI richer, and more fertile. It also allows us to move the discussion towards reflections on how to delimit the ecosystems, rather than looking at how to solve specific problems in specific parts”, says Bernd Stahl.

According to Bernd Stahl, this approach it not a cure for all that ails the AI ethics discussion. However, he claims it offers a novel way of thinking that is both complementary and compatible with the existing discourse, and that can help promote responsibility in AI ecosystems.

Want to read the full paper? Published Open Access in Scientific Reports: Stahl, B.C. Embedding responsibility in intelligent systems: from AI ethics to responsible AI ecosystems. Sci Rep 13, 7586 (2023). https://doi.org/10.1038/s41598-023-34622-w