



Safe Face Pledge Inactive



Safe Face Pledge Inactive

Safe Face Pledge - Inactive
Pledge to mitigate abuse of facial analysis technology—any automated system that evaluates human faces or heads—by adopting 4 commitments
1Show
Value for Human Life, Dignity, and Rights
2Address
Harmful Bias
3Facilitate
Transparency
4Embed
Commitments into Business Practices
Commitment One: Show Value for Human Life, Dignity, and Rights
Signatories of the Safe Face Pledge agree to:
1Do not contribute to applications that risk human life
By acknowledging that decisions that foreseeably increase the risk to human life are too dangerous for artificial intelligence, and by refraining from selling or providing facial analysis technologies to locate or identify targets in operations where lethal force may be used or is contemplated.
2Do not facilitate secret and discriminatory government surveillance
By acknowledging the right of the public to understand whether and how facial analysis technologies are used by the government. By refraining from knowingly selling to the government any products and services that are not subject to public scrutiny, inspection, and oversight.
3Mitigate law enforcement abuse
By acknowledging the right of the public to control whether and how facial analysis technologies are used in local, state or federal law enforcement including immigration agencies. By refraining from selling to law enforcement any products and services for purposes of enforcing criminal law unless a governing legislative body has explicitly and publicly considered all potential harms and authorized use of the technology through statute or ordinance.
4Ensure your rules are being followed
By acknowledging a responsibility over how facial analysis technology is used and that ignoring a customer’s implementation or use of the technology can risk harming community members. By adopting internal “know your customer” policies and procedures to ensure as best you can that your products are not being used for secret government surveillance.
Commitment Two: Address Harmful Bias
Signatories of the Safe Face Pledge agree to:
1Implement internal bias evaluation processes and support independent evaluation
By adopting internal systems to evaluate the performance of your facial analysis products and services and also facilitating independent research in order to identify and mitigate harmful bias in the design, development, deployment, and use of these systems.
2Submit models on the market for benchmark evaluation where available
By submitting to ongoing external evaluation to the available standards bodies that benchmark facial analysis technology in the countries of operation.
Commitment Three: Facilitate Transparency
Signatories of the Safe Face Pledge agree to:
1Increase public awareness of facial analysis technology use
By publishing accessible information on how facial analysis technologies are sold and used, including the types of entities they are sold to and any safeguards taken to mitigate misuse and risks. By proactively making a public explanation of how the systems works in clear and simple terms so that the people can understand how they work.
2Enable external analysis of facial analysis technology on the market
By publishing accessible descriptions and results of national or international benchmark performance evaluations, where available, including demographic information about the training data and collection processes.
Commitment Four: Embed Safe Face Pledge into Business Practices
Signatories of the Safe Face Pledge agree to
1Modify legal documents to reflect value for human life, dignity, and rights.
By updating vendor contracts, partner agreements, and terms of service to reflect commitments associated with lethal, secret government surveillance, and law enforcement applications, including terms that obligate compliance with these commitments and requiring the cessation of use of the technology in the event of any discovered non-compliance.
2Engage with stakeholders
By proactively reaching out to affected communities, as well as experts in data ethics, civil rights, civil liberties, racial and gender justice, privacy, and policing for consultation or cases for cessation regarding the development and deployment of facial analysis technologies.
3Provide details of Safe Face Pledge implementation
By publishing a statement detailing how your organization complies with the commitments set forth in this pledge within six months of signing.
Sunset Project of the Algorithmic Justice League and the Center on Privacy & Technology at Georgetown Law