Beginner
Blueprint for an AI Bill of Rights
Office of Science and Technology Policy · The White House · 2022
Read the original paperPlain-English Summary
A non-binding framework from the White House identifying five principles to guide the design, use, and deployment of AI systems: safe and effective systems, algorithmic discrimination protections, data privacy, notice and explanation, and human alternatives.
PolicyGovernanceRegulationEthicsPrivacy
published: false
Why This Paper Matters
While not legally binding, this document represents one of the most prominent attempts by a government to articulate what rights people should have in relation to AI systems. It has influenced subsequent legislative proposals in the US and internationally.
Key Principles
- Safe and effective systems: You should be protected from unsafe or ineffective AI systems.
- Algorithmic discrimination protections: AI systems should not contribute to unjustified different treatment based on protected characteristics.
- Data privacy: You should be protected from abusive data practices and have agency over how your data is used.
- Notice and explanation: You should know when an AI system is being used and understand how it affects you.
- Human alternatives: You should be able to opt out and have access to a human alternative when appropriate.
Discussion Questions
- Should these principles be made legally enforceable? What would that look like?
- How do these principles compare to the EU AI Act's approach?
- Are there important rights missing from this framework?