October 8, 2019

Ethical Standards: AI Trust Building

By: Larry Bridgesmith

Every time we buckle a child in an automobile safety seat, use a dry erase marker on a white board or place our trust in food safety management before we take that first bite, non-governmental international organizations probably have issued a set of standards which define best practices for that product or service.

The grandparent of standard setting organizations is the International Standards Organization (ISO). ISO had its origins in 1928 and was formally recognized by the United Nations in 1947 as an authorized UN advisory group. Over the time of its existence, ISO has approved over 22,000 standards using a complex consensus based process involving stakeholders who have an interest in the design, production, marketing or use of the subject goods and services.

ISO provides a short introduction into how these non-government standards impact those of us who depend on the quality, effectiveness and safety of the products on which we depend.

ISO is but one of many standards setting organizations globally. In the technical industry alone there are over 80 international standards organizations which verify the best practices in developing and producing technology related goods and services.

A prominent engineering and technology organization with its own standards setting body is the Institute of Electrical and Electronics Engineers and its Standards Association (IEEE). The IEEE has just published its first set of standards governing ethical Autonomous and Intelligent Systems (A/IS) design, development and deployment. ETHICALLY ALIGNED DESIGN: A Vision for Prioritizing Human Well-being with Autonomous and Intelligent Systems resulted from four years of intense work by hundreds of experts in engineering, machine learning, ethics, law and public policy.

A/IS holds tremendous promise for advancing and improving the human condition in health care, justice, education and economic opportunity for all the world’s population. A/IS applications have also been cited for bias, exploitation, unnecessary risk and manipulation of the user base.

IEEE developed its P7000 standards to promulgate ethical design principles applicable to technology design and development. Ethically Aligned Design focuses on the policies, principles, legal and ethical boundaries of A/IS from the perspective and the purpose of human benefit.

The foundations of Ethically Aligned Design (EAD) are its three pillars:

Based on these pillars, the eight principles of EAD are:

Application of the three pillars to the eight principles leads to the following matrix:

Working within these intersections can lead to trustworthy applications of A/IS when designers and developers commit to adhere and subject themselves to transparency and accountability in achieving these objectives.

This highly collaborative and self-governing approach to advanced technology development and implementation can transform the black box approach to AI which so many critics correctly deride. The IEEE EAD project provides a roadmap for the ethical work AI engineers, tech companies and users of these emerging technologies can cite to engender trust and value to the work being done to promote the values of a global system of data driven justice, education, health care and economics.

Specifically, the IEEE publication addresses these critical concerns in its format and content:

Almost all the concerns raised by the critics and the legitimate concerns directed at AI today can be addressed by the recommendations and principles espoused and articulated by EAD.

Based on the excellent work of IEEE and its contributors it is clearly time to move forward. Building solutions within a context of responsible behavior and value based goals to achieve the best outcomes in serving human needs through digital data is an achievable goal.

Why does this matter? It’s only the beginning!

If the transformative work being done in A/IS is to be trusted by the global population it seeks to serve, standard setting is a process by which the current state can be guided by informed multi-disciplinary recommendations for achieving a preferred future state. Unlike government regulation which tends to be poorly informed, politically driven and reactive, self-governing standards can anticipate, define best practices and adjust as circumstances change in a constant state of process improvement.

The IEEE ethical guidelines serve as beginning point and an explicit call for global participation in developing relevant standards for the many components of trustworthy A/IS development.

Nicolas Economou, chair of the Law Committee for this IEEE initiative outlines next steps for creating trustworthy A/IS technology. As he states:

“It’s just like we do in data security. You don’t use a data centre unless it’s ISO certified, for example. This makes procurement very easy and trustworthy. That’s the idea,’ he added.

“In terms of roll-out, he explained: ‘This is not going to come out with 100 standards of certification all in one year – it’s going to be progressive. There’s still a lot of work to do, but we are very actively involved in that, because we think it’s essential work for the justice system.

“Again, I want to be a little cautious because it may start with one set of certificates for example on accountability, on transparency, to show that something is transparent or that accountability can be traced. Things like that will be our first efforts,”

For those of us in the legal sector, this process is exactly what the American Bar Association appealed to when its Board of Governors approved Resolution Number 112 in August 2019. The ethics and applications of A/IS technology require global engagement by all disciplines impacted by these amazing technologies.

The time to stand on the sidelines and observe has passed.

Mr. Bridgesmith has over 40 years experience in legal professional services and numerous business ventures involving digital technologies. He has represented, trained, and consulted with organizations large and small in most industries. He is currently a Managing Partner of Accelerate InSite with a focus on AI Strategies and Implementation.