The White House’s Office of Science and Technology Policy (OSTP) released a blueprint for an Artificial Intelligence (AI) Bill of Rights, a set of five principles and associated practices to help guide the design, use and deployment of automated system, according to a release.
The goal of the document and its accompanying handbook is to provide a roadmap for governments, corporations, academics, researchers and other stakeholders to better develop technologies to help rather than harm millions of consumers in housing, financial services, education and other public sectors.
“Technology is the new civil rights frontier,” President and CEO of the National Fair Housing Alliance (NFHA) Lisa Rice said in a release. “We are learning more each day about how existing technologies harm people and communities. This plan will serve as a guide that encourages developers and marketers of AI tools to search for least discriminatory alternative (LDA) solutions in credit scoring, underwriting, pricing, tenant screening, health management, employee screening and other systems.
“It will also continue to position the United States as a global leader in advancing policies and techniques for the development of fair, transparent, explainable, responsible AI.”
The need for such a document arose from recognized issues with AI and the algorithms used. Tools created with this technology to assist with hiring and credit decisions have been found to reproduce undesired inequities or even create harmful bias and discrimination. When it comes to housing, NFHA stated race-based laws and policies, segregation, private prejudices, real estate agent steering, bank redlining, appraisal bias and restrictive zoning ordinances have contributed to biased data pools, resulting in discriminatory algorithms.
“Already, communities of color are facing disproportionate adverse effects from the rapid proliferation of AI technologies in the United States,” David Brody, managing attorney of the digital justice initiative at the Lawyers’ Committee, said. “The Bill of Rights establishes a new generation of fair information practices. Automated decision-making systems are used to determine access to and eligibility for essential needs such as housing, employment, credit, healthcare, education, insurance and public accommodations. Despite claims that these AI technologies are ‘advanced,’ they cannot distinguish between acceptable patterns of data and those that are the product of historic and repeated inequities.”
However, like all tools, AI cannot be labeled as “bad” or “dangerous”; rather, it is how it is used that determines its impact. The blueprint’s goal is to protect consumers and promote equity in areas where automated systems can meaningfully impact the public’s rights, opportunities, or access to critical needs.
“If broadly implemented, the plan can reverse the underrepresentation of Black, Latino, Asian, Native and other people of color in housing, economic, health, education, employment and other data and it can mitigate the ability of algorithmic systems to perpetuate injustices,” NFHA Chief Tech Equity Officer Michael Akinwumi said.
The White House’s blueprint lays out five core protections to which consumers should be entitled: safe and effective systems, algorithmic discrimination protections, data privacy, notice and explanation, and alternative options.
In short, consumers should:
- Be protected from unsafe or ineffective systems.
- Be protected from discrimination by algorithms, and systems should be used and designed in an equitable way.
- Be protected from abusive data practices via built-in protections.
- Have agency in how data about them is used.
- Be informed an automated system is being used and understand how and why it contributes to outcomes impacting the consumer.
- Be able to opt out, where appropriate, and have access to a person who can consider and remedy issues in a timely manner.
“Privacy rights are civil rights,” President and Executive Director of the Lawyers’ Committee Damon Hewitt said. “As our lives are increasingly spent online, it is more important than ever that racial equity is advanced in the digital world. The AI Bill of Rights establishes important principles for data use by both the government and the private sector. Even with good intentions, big data technologies can reproduce and reinforce patterns of discrimination. We need safeguards against algorithmic bias and other forms of digital discrimination to ensure the data used to inform artificial intelligence is unbiased and that the results are equitable.”
ReNika Moore, director of the American Civil Liberties Union’s Racial Justice Program, said it is critical for the Biden administration to make the AI Bill of Rights blueprint a reality.
“We are encouraged to see the administration announce some concrete steps by agencies and we urge the administration to commit to more concrete enforcement efforts, regulation, and oversight across the federal government to make the rights in the blueprint real,” she said. “Equally important, there should be no loopholes or carve-outs for these protections. The federal government must enforce these rights in the law enforcement and national security contexts, where the harms to people from automated systems are well-documented and more common and severe for marginalized groups, and where people’s liberty and due process rights are routinely at stake.”
Efforts to regulate AI to promote fairness and equity also are being made by the Consumer Financial Protection Bureau, the Federal Trade Commission and the U.S. Department of Housing and Urban Development (HUD). Specifically, HUD announced it plans to release guidance related to tenant screening algorithms. Other federal agencies that regulate mortgage financing intend to implement a nondiscrimination quality control standard for automated valuation models to combat biased data from past housing discrimination.
“To fully implement the principles in this new AI Bill of Rights, congressional action is needed,” Brody said. “Congress must pass The American Data Privacy and Protection Act and build out our civil rights infrastructure to fight discrimination in a data-driven economy.”