What happened
The IEEE Standards Association is marking ten years since the first draft of Ethically Aligned Design, its framework for building ethics into AI and autonomous systems, IEEE Spectrum reported. For the anniversary, the association has launched a website, a blog and a five-part podcast series. The podcast began on 21 October, with new episodes released every two weeks.
The framework underpins the IEEE 7000 family of standards, which now spans 29 projects, 11 of them published, and more are still being added. The published standards cover areas such as transparency of AI systems, data privacy, algorithmic bias, effects on human well-being, machine-readable personal privacy terms and systems that imitate human empathy.
Alongside the standards, IEEE SA runs CertifAIEd, which offers one certification for individuals and another for products. A CertifAIEd assessor checks whether an AI system meets the criteria in the IEEE Ontological Specifications, and the IEEE Conformity Assessment Program can then grant a certification mark. The association has also recently introduced a Responsible AI for Children certification for AI products aimed at young users.
Context
IEEE SA started its global initiative on ethics in autonomous and intelligent systems in 2016 and published the first Ethically Aligned Design report that year. The report set out eight principles, including human rights, well-being, transparency and accountability. A second version followed in 2017, and by 2019 the effort had more than 400 contributors. A related IEEE industry group on AI and human flourishing now has more than 200 members.
O briefing
A semana em tecnologia, em um só e-mail.
As reportagens que valem o seu tempo, toda semana. Cancele quando quiser.
