The White House Office of Science and Technology Policy (OSTP) released its Blueprint for an AI Bill of Rights and, thankfully, it doesn’t give AI any rights. This (nonbinding) document is meant to provide a best-practice guide for AI system designers and frames those systems’ use as a civil rights issue. The OSTP focused on consumer AI interactions, excluding many internal business AI applications.
The blueprint outlines five principles, along with detailed descriptions of problematic AI uses to illustrate these points:
(1) Safe and Effective Systems: focusing on systems with intentional or foreseeable harmful effects
(2) Algorithmic Discrimination Protections: focusing on intentional or unintentional discrimination between groups from AI systems
(3) Data Privacy: focusing on consumer choice and control over their personal data
(4) Notice and Explanation: focusing on clear, timely and accessible disclosures to consumers
(5) Human Alternatives, Consideration & Fallback: focusing on the consumer’s right to opt-out of AI data processing
The White House also announced that several agencies will implement related actions and guidance regarding their AI systems, including new procurement policies. Overall, this document is a welcome (even if ultimately toothless) attempt to set AI industry standards that value ethics and human dignity. This white paper is only the beginning of a legislative trend toward more regulation and restrictions on the use of AI and the ethical considerations when using such technology.