Paul Lewis, director of ethics and business conduct for the intelligence and security sector at BAE Systems Inc., wrote in a commentary published Tuesday on Federal News Network that there are several issues the Department of Defense’s Joint Artificial Intelligence Center should address when implementing the five ethical principles for AI and one is clarifying the terms used in the principles.
Lewis said Alka Patel, head of AI ethics policy team at JAIC, and her team should clarify the terms “appropriate” in the first principle of responsibility and “unintended bias” for the equity principle, among others.
Another factor to consider is determining the scope of AI’s human control in military application.
Lewis said JAIC should also examine the scope of human accountability for decision making with regard to the use of AI and address the issue surrounding “moral machines.”
He also cited some of the initial steps Patel and her team have started to implement the five AI ethical principles. Those include the establishment of the Responsible AI Champions pilot program and formation of a Data Governance Council within the U.S. government.