Dr. Keng Siau, chair and professor of business and information technology and Weiyu Wang, a 2019 Missouri S&T master’s graduate in information science and technology, propose a new dimension in the ethics of artificial intelligence (AI) in a conceptual paper.

The paper, titled “Artificial Intelligence (AI) Ethics: Ethics of AI and Ethical AI,” is published in the April-June 2020 issue of the Journal of Database Management.

The researchers suggest that not only should AI agents not harm human beings, but ethical AI machines should also behave ethically in their interactions with other AI agents.

“This aspect is increasingly important in AI development because strong machines have an intellectual capability that functionally equals or exceeds that of humans,” Siau says. “Loosely speaking, strong AI machines have a mind of their own that can process and make independent decisions.”

The researchers distinguish the need for both “AI ethics,” the study of ethical issues in AI, from “ethical AI,” the AI agents that need to perform ethically based on moral values, principles, policies and regulations that define how the agents should interact with humans and society, as well as with other AI agents.

For example, unethical uses of AI that could be prevented with more oversight include voice fraud, which rose 350% between 2013 and 2017, superimposition of false content on source images or videos for political purposes, and false-labeling of high-risk individuals in the criminal justice system based on ethnicity.

To date, the existing AI ethical frameworks vary from institution to institution — from the United Nations Educational, Scientific, and Cultural Organization (UNESCO) to Google to Accenture, and more. The researchers propose an overarching ethical AI framework, noting that across the globe, today’s AI makers and researchers are likely to pay more attention to performance metrics such as safety, reliability, usability and customer satisfaction than more nebulous concepts, like ethics, due to the intense competition between companies and nations.

Keng and Wang conclude that the future of humanity may depend on the correct development and strong enforcement of AI ethics, and that stakeholders need to make ethics and morality central to AI’s revolution and evolution.

Keng is editor-in-chief of the Journal of Database Management, a quarterly peer-reviewed academic journal covering database and information management published by IGI Global.