Is it time for a Bill of Rights to keep AI in its place?
The White House wants to create a what it is calling a bill of rights protecting people from harm by AI.
Its first step has been to ask for the input of people and organizations who have experience working with biometrics and biometrics-based systems, including those designed to recognize emotions.
Officials in the Office of Science and Technology Policy are asking the question many have been pondering for a decade: How can AI’s power be constrained when it comes to individuals and society?
Eric Lander and Alondra Nelson, director and deputy director, respectively, of the office, made an unsettling analogy about the topic in an essay in culture-of-tech publication Wired.
The Bill of Rights was needed when the United States was founded to protect people from a powerful national government, they wrote. People now need a bulwark against those who wield AI, and, presumably, against artificial general intelligence that some feel is inevitable.
That comparison seems to assume that AI could grow as powerful as the national leaders who are only beginning to take it seriously.
For their parts, Lander and Nelson, have decided their first bite of the apple will be a request for information from pretty much anyone who has had experience with development and operation of biometric recognition systems.
They are focusing their search on software that infers the mental and emotional states of individuals, identity verification and the identification of individuals. That is a pretty big first bite.
The latter two topics are controversial (See: Clearview AI) but still more demonstrably accurate than emotion recognition, or affective computing, which people in the industry and out say is impossible to do and may always be impossible.
That has not stopped vendors and their consultants from claiming successful products and brighter futures.
AI | biometric identification | biometrics | emotion recognition | identity verification | legislation | regulation | RFI | U.S. Government