The AI Bill of Rights: Defining fairness and privacy by design

0
27

The White House Office of Science and Technology Policy set its expectations of automated system development based on five principles guiding design, use and deployment in its “Blueprint for an AI Bill of Rights.”

WHY IT MATTERS
To protect the American public in the age of artificial intelligence, the “Blueprint for an AI Bill of Rights” is a guide that protects people from threats and a framework that defines guardrails on technology to reinforce civil rights, civil liberties and privacy, and equal opportunities ensuring access to critical resources and services.

While financial, public safety, social services, government benefits, and goods and services are named, healthcare AI is called out first.

“Too often these tools are used to limit our opportunities and prevent our access to critical resources or services. These problems are well-documented. In America and around the world, systems supposed to help with patient care have proven unsafe, ineffective or biased,” according to the White House.

The blueprint and its accompanying handbook are a response to public concern about the use of AI to make decisions, and they have been developed by researchers, technologists, advocates, journalists and policymakers, according to the announcement.

The “From Principles to Practice” handbook includes detailed steps toward actualizing the following principles in the technological design process:

  • Safe and effective systems.

  • Algorithmic discrimination protection.

  • Data privacy.

  • Notice and explanation.

  • Human alternatives, consideration and fallback.

Each principle is defined so that automated systems “can meaningfully impact the public’s rights, opportunities or access to critical needs,” and there are references to President Joe Biden’s remarks and executive orders throughout, including on advancing racial equity for the underserved, the Supreme Court’s decision to overturn Roe v. Wade, and more.

Footnotes to big data reports – including egregious justice errors based on bad facial recognition matches and racial bias, education redlining, biased hiring algorithms and labor recruitment tools, flawed population health AI and more – can be found throughout the document.

The blueprint directs AI developers to consult with diverse communities, stakeholders and domain experts to identify risks and potential impacts in order to develop safe, effective systems.

“In order to ensure that an automated system is safe and effective, it should include safeguards to protect the public from harm in a proactive and ongoing manner; avoid use of data inappropriate for or irrelevant to the task at hand, including reuse that could cause compounded harm; and demonstrate the safety and effectiveness of the system,” according to a specific prescription for what should be expected of automated systems.

The blueprint also details consultation, risk identification and mitigation, ongoing human-led monitoring for the lifespan of deployed automated systems, and oversight responsibilities by AI system owners.

“Those holding this responsibility should be made aware of any use cases with the potential for meaningful impact on people’s rights, opportunities or access as determined based on risk identification procedures” – and when they are, “responsibility should rest high enough in the organization that decisions about resources, mitigation, incident response and potential rollback can be made promptly, with sufficient weight given to risk mitigation objectives against competing concerns,” according to the blueprint.

Impact on rights would of course include violations of patient privacy rights, and the blueprint calls for enhanced protections and restrictions for data across sensitive domains, like healthcare.

The practice section calls for privacy by design and by default where “data collection should be limited in scope, with specific, narrow identified goals, to avoid ‘mission creep.’ Anticipated data collection should be determined to be strictly necessary to the identified goals and should be minimized as much as possible.”

Additionally, the blueprint states, “Data collected based on these identified goals and for a specific context should not be used in a different context without assessing for new privacy risks and implementing appropriate mitigation measures, which may include express consent.”

Requests from members of the public about their data being used in a system should be met with a response by the entity responsible for its development, including a report on the data it has collected or stored about them and an explanation of how it is used.

THE LARGER TREND
AI products and services have the potential to determine who gets what form of medical care and when, for example, so the stakes are high when algorithms are deployed in healthcare, putting trust on shaky ground.

Algorithmic bias with respect to race, gender and other variables has raised concerns about the downstream effects of these models and has spawned efforts to drive evidence-based AI development in the healthcare space.

However, all data is biased, according to Dr. Sanjiv M. Narayan, co-director of the Stanford Arrhythmia Center, director of its Atrial Fibrillation Program and professor of medicine at Stanford University School of Medicine.

“There are multiple approaches to eliminate bias in AI, and none are foolproof. These range from approaches to formulate an application so that it is relatively free of bias, to collecting data in a relatively unbiased way, to designing mathematical algorithms to minimize bias,” he told Healthcare IT News in a discussion last November about how biases arise in AI and how to improve decision support.

While Wall Street Journal readers will note that many industry leaders responded with concern that the White House “Blueprint for an AI Bill of Rights” will lead to regulations that choke development, USA Today reported that Biden was on hand in New York this week as IBM announced $20 billion in investments for research and development and manufacturing, including AI and quantum computing.

“It’s here now where the Hudson Valley could become the epicenter of the future of quantum computing, the most advanced and fastest computing ever, ever seen in the world,” Biden said.

ON THE RECORD
“Tracking and monitoring technologies, personal tracking devices and our extensive data footprints are used and misused more than ever before; as such, the protections afforded by current legal guidelines may be inadequate,” the White House indicated in the blueprint’s “Extra Protections for Data Related to Sensitive Domains” section.

“The American public deserves assurances that data related to such sensitive domains is protected and used appropriately and only in narrowly defined contexts with clear benefits to the individual and/or society.”

Andrea Fox is senior editor of Healthcare IT News.
Email: afox@himss.org

Healthcare IT News is a HIMSS publication.

Source