Fujitsu, an information and communication technology firm, recently revealed new resource toolkit that guides developers in assessing risk related to artificial intelligence (AI) systems regarding international AI ethics guidelines. Fujitsu provides services for free with the aim of promoting safe and secure AI systems.
A newly designed approach for defining ethical needs in AI ethics guidelines written in natural language, as well as for applying ethical requirements to actual AI systems, is included in the toolkit.
Fujitsu aims to avoid misunderstandings and potential risks caused by interpretations of guidelines by providing AI system developers and operators with new tools for thoroughly identifying and preventing potential ethical issues early in the development process, in accordance with international best practices.
Dr. Christoph Lutge of the Technical University of Munich, a leading authority in the research of responsible AI and business ethics, comments, “In Europe, there is a growing debate about AI regulations, and one of the key issues is how to close the gap between principles and practices, or “what” and “how.” I believe that the results of this research are very significant in that they enable us to practice based on principles. I would also like to express my deep appreciation for the decision to open up the research results and stimulate discussion worldwide.”
An expanded version of the toolkit will be released by end of 2022 by working in collaboration with government agencies, private companies, and leading researchers to improve and furnish the methodology.
Fujitsu has conducted studies of past AI-related events recorded in the AI Incident Database of the international consortium Partnership on AI in developing this new toolkit and guidance for developers.
This method led to the conclusion that the flow of information (“interactions”) between discrete parts inside an AI system, as well as between an AI system and its users and other stakeholders, can be contextualized with ethical considerations relating to AI systems.
Fujitsu successfully created an evaluation approach based on these findings to systematically identify important ethical concerns connected to AI systems, enabling them to produce AI ethics models that can clarify the interpretation of AI ethics standards.