Principles in Practice: Establishing a governance framework for beneficial Autonomous and Intelligent Systems

Principles in Practice: Establishing a governance framework for beneficial Autonomous and Intelligent Systems
Like

This article is part of the Forum Network series on Digitalisation and reflects on discussions at OECD Forum 2019.

With the mission to advance technology for humanity, the societal and ethical aspects of technology have long since been part of IEEE’s DNA. In 2015 we established The IEEE Global Initiative on Ethics of Autonomous and Intelligent Systems to ensure everyone designing and developing A/I systems (a term that includes artificial intelligence systems, robotics and autonomous systems) is educated, trained, and empowered to prioritise ethical considerations so that these technologies are advanced for the benefit of humanity. Since then, we have introduced a variety of other platforms for open, multidisciplinary conversation and consensus building related to societal and ethical aspects of these technologies.


Notably, this work includes engaging with stakeholders from around the globe, and sharing our expertise with municipalities and national governments as well as with international organisations. In particular, IEEE co-operated with the Organisation for Economic Co-operation and Development (OECD) by contributing to a set of intergovernmental principles through the OECD AI Experts Group. There is a significant affinity between IEEE’s work and the OECD Principles on AI, specifically regarding the need for such systems to primarily serve human well-being through inclusive and sustainable growth; to respect human-centred values and fairness; and to be robust, safe and dependable, including through transparency, explainability and accountability.


There is a very good reason for significant efforts and investments by IEEE in this area. A/I systems and technologies hold great promise to benefit our societies but also present potential new social, legal and ethical challenges. Systemic cyber-physical risks, the possible exacerbation of power and wealth concentration and issues of data transparency, ownership and agency are some of the concerns that continue to surface in the public debate. Due to the nature of the concerns, they cannot be dismissed or handled as an afterthought. Key stakeholders including legislators, regulators and the technical community and industry should assume their responsibility and do their best – not just the minimum – to address them upfront. For IEEE’s audiences, which include developers and operators of A/I systems, it is important to build and maintain awareness of some critical issues such as the need for transparency, accountability and lack of bias. In addition, these issues have to be understood in their context; many questions in the health sector will be different to those in the financial or legal sectors.

In an almost four-year iterative process, the open, collaborative and empowered community of The IEEE Global Initiative grew to 2,500 members and produced the foundational work of Ethically Aligned Design (EAD). This freely available report provides a rigorous, in-depth analysis of key issues and a set of general principles that should be valid independently of the field of application, followed by high-level recommendations. The latter are specifically designed to serve as guidance for industry and practitioners to establish standards and certification for the design, manufacture and use of such systems.

IEEE Ethically Aligned Design Treatise First Edition
Image: IEEE

As mentioned above, the principles and recommendations of EAD are well aligned and complementary to the OECD AI Principles adopted by the OECD Ministerial meeting as part of OECD Week 2019. Their propagation, also through the OECD mechanisms, will help inform the evolution of regulatory, legal and other environments that will ensure future development of A/I systems is performed with appropriate care and in alignment with societal values and ethical principles.


Ultimately, in order to develop the level of trust between people and technology required for A/I systems to be truly beneficial to our daily lives, principles and recommendations need to move into practice. At IEEE, efforts to this end include the development of voluntary, bottom-up consensus standards, exemplified through the flagship IEEE P7000™ standards series and the development of certification processes to help people recognise and distinguish systems that comply with these standards. We are also placing emphasis on the needs of vulnerable groups, in particular by supporting the design of systems that respect and protect the rights of children in the Internet and A/I systems era. As part of its new principles, the OECD will be establishing the OECD AI Policy Observatory, an online hub that will provide resources for public policy topics, policies, initiatives and metrics. IEEE stands ready to contribute to the Observatory, and we congratulate the OECD on its achievements and the ongoing work.

OECD Principles on Artificial Intelligence

Continue the conversation and help us co-create the agenda

All of the discussions you have on the Forum Network inform our thinking for the OECD Forum event each year – join to respond to Konstantinos' article and comment to help us co-create the agenda

Related Topics

Artificial Intelligence OECD Forum 2019

Find out more about OECD Forum 2019: World in EMotion

OECD Forum 2019: World in EMotion

Banner image: Nathan Dumlao on Unsplash

Please sign in

If you are a registered user on The OECD Forum Network, please sign in