Fairness OS
Fairness OS is a technology development portfolio concerned with how AI treats people. Its five fields address robots, AI-assisted decisions, changing AI behavior, experiences combining text, images and audio, and actions performed by AI.
Discuss this technology
Fairness OS / PUBLIC BENEFITS
Continuing attention to fairness
Responsible adoption
Keep the people affected by AI at the centre of fairness discussions.
WHAT THIS MEANSDesigned around
Designed around
the people it serves.
People in different settings
Recognizes that fair treatment has different implications in a physical interaction, a digital decision or a creative experience.
Continuing attention to fairness
Addresses the need to consider fairness throughout an AI service’s use as its behavior and application change.
Responsible adoption
Gives partners a way to discuss the fairness needs of their intended AI application and the people it may affect.
FIVE DISTINCT FIELDSDifferent settings.
Different settings.
A shared concern for people.
R1
Physical Action Fairness
Keep fair treatment of people in view when physical systems act around them.
R2Causal Fairness
Put fair treatment of people at the centre of AI-assisted decisions.
R3Runtime Fairness & Drift Correction
Keep fairness relevant as an AI service changes.
R4Multimodal Fusion Fairness
Consider fair treatment across the different forms of content an AI experience uses.
R5Agent Fairness, Execution & Outcome Feedback
Make fair treatment and responsibility part of the discussion about actions performed by AI.
APPLICATIONSWhere it could
Where it could
make a difference.
- Robots and AI services used around people.
- AI-assisted decisions and experiences combining text, images or audio.
- Organizations considering fairness in changing AI services and AI actions.
Fairness goals, affected people and validation requirements need to be considered for each application. Responsible human judgment remains part of that discussion.
AI SAFETY GUIDESWhen AI enters
When AI enters
the real world.
Practical answers about Physical AI and Embedded AI: risks, safety, security and the questions to ask before deployment.
Physical AI safety.
For teams evaluating robots, autonomous machines and AI systems that act around people.
Embedded AI safety.
For teams evaluating smart devices, industrial equipment, sensors and on-device AI.
START A CONVERSATION
Partner with usDiscuss Fairness OS.
Tell us your application, the people involved and the outcome you want to work toward.