AI Bill of Rights: Defining fairness and privacy by design

The White House Office of Science and Technology Policy sets out its expectations for automated system development based on five principles that guide design, use, and implementation in the “blueprint for the AI ​​Bill of Rights.” .

WHY IT IMPORTANT
To protect the American public in the age of artificial intelligence, the “Blueprints for the AI ​​Bill of Rights” is a guide to protecting people from threats and a framework that identifies public barriers. technology to strengthen civil rights, civil liberties and privacy, and equality of opportunities ensuring access to vital resources and services.

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

“These tools are often used to limit our opportunities and prevent us from accessing critical resources or services. These issues are well-documented. In the US and around the world. systems that are supposed to help care for patients have been shown to be unsafe, ineffective, or biased,” according to the White House.

The blueprint and accompanying handbook are responses to public concern about using AI for decision-making, and they have been developed by researchers, technologists, advocates, newspapers and policymakers, according to the announcement.

The “From Principles to Practice” handbook includes detailed steps for realizing the following principles in the technology design process:

• Safe and efficient system.

• Algorithmic discriminant protection.

• Data security.

• Notice and explanation.

• Human alternatives, considerations and contingencies.

Each principle is defined so that automated systems “can meaningfully impact the rights, opportunities, or accessibility of important public needs” and includes references to comments by the public. President Joe Biden and cross-cutting executive orders, including promoting racial justice for the underserved, the Supreme Court’s overturned Roe v. Wade decision, and more.

Captions for big data reports – including severe justice errors based on bad facial recognition matches and racial bias, educational redirects, biased recruitment algorithms, and recruitment tools, population health AI with cons and more – can be found throughout the document.

The Blueprint guides AI developers in consultation with diverse communities, stakeholders, and industry experts to identify potential risks and impacts in order to develop innovative solutions. Safe and efficient system.

“To ensure that an automated system is safe and effective, it must include safeguards to proactively and continuously protect the public from harm; avoid inappropriate use of data or unrelated to the task at hand, including reuse that could cause complex harm; and demonstrate the safety and effectiveness of the system, “under a specific prescription for what should be expected of automated systems.

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

“Those who hold this responsibility should be made aware of any use cases that have the potential to have a meaningful impact on people’s rights, opportunities, or accessibility as determined based on validation procedures.” risk assessment” – and when it does, “responsibility should be placed high enough within the organization Resource, mitigation, incident response and payback decisions can be made in a timely manner, with sufficient weight.” quantity given for the purposes of mitigating risk against competitive concerns”, according to the blueprint.

The impact on rights would, of course, include a breach of patient privacy, and the blueprint calls for enhanced protections and restrictions on data across sensitive areas, such as healthcare. health care.

The practice section claims privacy by design and by default, where “data collection should be limited in scope, with specific, narrowly defined goals, to avoid ‘errors in mission’.” Expected data collection should be determined to be absolutely necessary for the defined objectives and should be minimized as much as possible. “

In addition, the blueprint states, “Data collected against these defined objectives and for a particular context should not be used in another context without assessing new risks to privacy and take appropriate mitigation measures, which may include explicit consent.”

Requests from members of the public that their data is being used in the system must be met by the entity responsible for the development of the system, including reporting on the data it has collected. collect or archive about them and explain how it is used.

TREND TO BIGGER WOMAN
For example, AI products and services have the potential to determine who gets medical care in what form and when, so the stakes are high when algorithms are deployed in the healthcare sector. health care, put your trust in an unstable foundation.

Algorithmic bias towards race, gender, and other variables has raised concerns about the downstream effects of these models and has prompted efforts to advance evidence-based AI development. in the health care sector.

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

“There are many approaches to removing bias in AI, and none of them are perfect. These include approaches to building an application such that it is relatively bias-free, to collecting data in a relatively unbiased way, to designing mathematical algorithms to minimize bias,” he said. IT news about healthcare in a discussion last November about how bias arises in AI and how to improve decision support.

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

“This is where the Hudson Valley could become the future focal point of quantum computing, the most advanced and fastest computer ever seen in the world,” Biden said.

ON PROFILE
“Our tracking and surveillance technologies, personal tracking devices, and expanding data footprint are used and abused more than ever before; thus, the protections provided by applicable legal guidelines may not be sufficient,” the White House states in its blueprint “Additional Protections for Sensitive Domains Part data”.

“The American public deserves assurance that data relating to such sensitive areas is protected and used appropriately and only in narrowly defined contexts with clear benefits to individuals. individual and/or society.”