Inroads - December 2017

OPINION

creasingly vary from each other. In addition to presenting a lack of precision in legal terminology, AI also involves an ever-growing number of stakeholders such as clients, insurers, law firms, contractors and regulatory entities, who all have different—and often opposing—interests.

While some researchers believe that the implementation of abstract values into AI systems is feasible, other scholars vehemently disagree that such an approach is either feasible or desirable. Etzioni and Etzioni [ 2] provide compelling arguments to demonstrate this position. They assert that a significant part of the ethical challenges posed by AI-equipped devices can be addressed by law enforcement and the personal choices of the users. Personal choices governed by laws are examples of ethical decisions made collectively and enforced collectively. AI devices must be programmed to observe these legal constraints, which do not involve an ethical decision for the device. For other more nuanced decision-making, it will be very difficult to program a moral code into an AI device. Furthermore, attempting to automate ethical norms into AI systems will create questions about the social position that AI-entities occupy in society. The future of AI relies heavily on the quality and capabilities that scientific researchers and engineers can deliver with particular emphasis on the proven safety and reliability of systems as they interact with humans. Examples such as driverless cars, automated sentencing recommendations to judges, smart voting systems, and smart home appliances all make great case studies to discuss with students how ethics can be used to build or can be built into such systems.

As they stand now, the EU commission report and US NAIRD plan are orthogonal documents with each one emphasizing a different approach to dealing with the ethical concerns raised by AI systems. The EU report focuses on greater human responsibility while the US report emphasizes optimism in the power of technology to solve ethical problems. It may be that both approaches will be needed to regulate the legal and ethical implications of AI—

greater oversight of human decision makers as well as ethical design and regulation of AI empowered devices. In a subsequent article we will discuss more fully the concept of the “personhood” of robots as well as the liability and responsibility issues inherent in AI technologies. 

References

1. Devalaux, M. (Rapporteur). European Union Report with Recommendations to EU Commission on Civil Law Rules on Robotics, January 2017; http://www. europarl.europa.eu/sides/ getDoc.do?pubRef=-//EP// TEXT+REPORT+A8-2017-0005+0+DOC+XML+V0// EN. Accessed 22 August 2017.

2. Etzioni, A. and Etzioni, O. Incorporating Ethics into Artificial Intelligence. Journal of Ethics (online issue). March 7, 2017; https://doi.org/10.1007/s10892-017- 9252-2

3. Kemp, R. Legal Aspects of Artificial Intelligence (pp. 1-2, Rep.); http://www.kempitlaw.com/wp- content/uploads/2016/11/Legal-Aspects-of-AI-2017- Branding-.pdf. Accessed 22 August 2017.

4. Moynihan, D. The Rise Of The Machines? Mondaq Business Briefing. March 24, 2017.

5. Networking and Information Technology Research and Development Subcommittee.

The National Artificial Intelligence Research and Development Strategic Plan (October 2016), 26-7; https://www.nitrd.gov/news/national_ai_rd_ strategic_plan.aspx. Accessed 22 August 2017.

C. Dianne Martin

Professor Emeritus of Computer Science and Vice Provost for Faculty Affairs The George Washington University Rice Hall – Suite 813

2121 I Street, NW Washington, DC 20052 USA

dmartin@gwu.edu

Toma Taylor Makoundou

Graduate Student of Computer Science, George Washington University, Washington, DC, 20052 USA

m@gwmail.gwu.edu

DOI: 10.1145/3148541 Copyright held by authors.

ACM
Transactions on

Accessible Computing

;;;;;

This quarterly publication is a quarterly journal that publishes refereed articles addressing issues of computing as it impacts the lives of people with disabilities. The journal will be of particular interest to SIGACCESS members and delegrates to its affiliated conference (i.e., ASSETS), as well as other international accessibility conferences.