Algorithmic Discrimination Protections

"You should not face discrimination by algorithms and systems should be used and designed in an equitable way. " 

-- The White House

- Algorithmic Discrimination Protections

Algorithmic discrimination occurs when automated systems contribute to unjustified different treatment or impacts disfavoring people based on their race, color, ethnicity, sex (including pregnancy, childbirth, and related medical conditions, gender identity, intersex status, and sexual orientation), religion, age, national origin, disability, veteran status, genetic information, or any other classification protected by law.  

Depending on the specific circumstances, such algorithmic discrimination may violate legal protections. Designers, developers, and deployers of automated systems should take proactive and continuous measures to protect individuals and communities from algorithmic discrimination and to use and design systems in an equitable way.  

This protection should include proactive equity assessments as part of the system design, use of representative data and protection against proxies for demographic features, ensuring accessibility for people with disabilities in design and development, pre-deployment and ongoing disparity testing and mitigation, and clear organizational oversight.  

Independent evaluation and plain language reporting in the form of an algorithmic impact assessment, including disparity testing results and mitigation information, should be performed and made public whenever possible to confirm these protections. 

- The AI Bill of Rights

The AI Bill of Rights is a set of guidelines for the responsible design and use of artificial intelligence. The White House Office of Science and Technology Policy (OSTP) created the bill. The bill includes five principles: 

• People should be protected from systems deemed “unsafe or ineffective”.
• People shouldn't be discriminated against via algorithms.
• AI-driven systems should be made and used “in an equitable way”.
• People should have access to timely human consideration and remedy if an automated system fails or produces an error.
• People should be able to opt out of automated systems.

The bill was created amid a global push to establish more regulations to govern AI. The Biden Administration calls the guidelines “common sense protections to which all Americans should be entitled”.

[New York City, New York - Civil Engineering Discoveries]

- AI Legislation

Algorithmic discrimination can occur when algorithms inadvertently discriminate without looking at protected characteristics. For example, a facial recognition algorithm might be trained to recognize white people more easily than Black people. 

The Equal Employment Opportunity Commission (EEOC) issued guidance in 2022 about algorithmic bias in the workplace. State and local lawmakers are also drafting and enforcing restrictions on the discriminatory use of artificial intelligence in the workplace. 

The AI Bill of Rights includes a principle that calls for proactive equity assessments of algorithms and ongoing disparity mitigation. 

The District of Columbia reintroduced the Stop Discrimination by Algorithms Act of 2023 to prohibit algorithmic discrimination by employers. The AI Disclosure Act of 2023 would require a disclaimer on all material generated by artificial intelligence. 

To mitigate bias in machine learning models, you can: 

• Use a diverse and representative training dataset
• Remove sensitive variables
• Use bias mitigation techniques
• Regularly evaluate the model
• Use human oversight

Using demographic information in the design, development, or deployment of an automated system can lead to algorithmic discrimination.

[More to come ...]