Artificial intelligence technologies are today embedded in a range of U.S. industries and firms, including the health care, transportation, and manufacturing sectors, as well as such products and services as e-mail spam filtering, graphic imaging applications, and financial lending decision software. Public policy discussions are now focused on five public licensing proposals/models designed to regulate evolving AI technologies and best meet legal and ethical challenges: open source/no patent; fair, reasonable, and non-discriminatory (FRAND); evolving licensing system; public interest licensing; and compulsory licensing. After a careful analysis, the most promising public licensing policy model involves an evolving licensing system proposal. Given the rapid advances in recent AI technologies development, this licensing approach is adaptive to future commercial needs, technology advances, and public policy, safety, and ethical concerns. The paper also explores the nuances of needing to “thread-the-needle” correctly in the public–private regulatory policy approach and the economics of the “public interest”.
Hemphill et al. (Fri,) studied this question.