Opinion: An AI bill of rights

By BILL RYAN

NH Business Review

Published: 06-06-2023 4:08 PM

Bill Ryan writes about career, employment and economic topics from his home in North Sutton.

Often, it is difficult to separate living from working. Our personal lives and professions can become intertwined such that it can seem pointless to differentiate those aspects which are personal from professional. Such is the case when considering one of today’s hottest topics: the impact of artificial intelligence. Is AI going to sway our lives in general, or be mostly an employment issue? A fair prediction is that AI is going to change the landscapes of both our lives and of our work.

As citizens and as workers, we should have a strong say in what the influence of AI is going to be in our daily lives and on our jobs. The disruptive potential is too huge to leave AI development solely up to engineers and their corporate employers. If AI advancements are to be the result of free market innovation, then those of us who are future customers and recipients of its consequences should have the freedom to weigh in and heavily influence its maturation.

A practical way to approach this challenge is through the lens of individual rights.

Ever since the 17th century philosopher John Locke proposed the existence of fundamental natural rights, such as life, liberty and property, we Westerners have organized our social, political and economic institutions around the notion of personhood rights to both preserve and extend the enjoyment of our lives. We bestow upon ourselves the rights necessary to live fruitful lives free of destructive intrusion. Now is the time to apply these rights in the face of AI infiltration.

A useful place to ground a national debate about AI’s proliferation is with the Biden administration’s White House Office of Science and Technology Policy’s proposal, known as the Blueprint for an AI Bill of Rights. This is a thoughtful approach to identifying the key areas of contention in the planning, application and mobilization of AI-based automated systems.

Five principles are presented as foundational to designating what constitutes an AI Bill of Rights. To summarize:

• Safe and effective systems: An AI system should undergo input and testing from various sources to ensure its ability to deliver value free from the risk of malicious or unintended consequences. Humane industry standards and protective measures should apply, including the power to shut down harmful applications. Data usage is to be transparent, necessary and respectful of personal integrity.

Article continues after...

Yesterday's Most Read Articles

“It’s beautiful” – Eight people experiencing homelessness to move into Pleasant Street apartments
No deal. Laconia buyer misses deadline, state is out $21.5 million.
N.H. Educators voice overwhelming concerns over State Board of Education’s proposals on minimum standards for public schools
Matt Fisk will serve as next principal of Bow High School
High schools: Coe-Brown softball wins 5th straight, Concord’s McDonald pitches first varsity win, Tide’s Doherty scores 100th career point
Construction of housing project in former Church to begin with parking dispute in the rearview

• Algorithmic discrimination protections: The biases, inequities and discriminatory practices of people should not migrate to automated systems. Indefensible digital treatment of people based on their individual differences is to be considered unjust. Legal protections of ordinary citizens and farsighted equity assessments of intended and unintended uses of systems should be crucial in the design and deployment of AI systems.

• Data privacy: This concern has been with us since the advent of Web 2.0. People should have ownership and agency over their data. The right to privacy is strong among free and independent people. This should be reflected in the automated systems they use. Exercising consent and having the ability to opt in and out of these systems with no restrictions should be inherent in their development.

• Notice and explanation: It should not take a computer science degree for ordinary users to understand what they are getting into with AI systems. Clear and unambiguous language that informs operators about system functionality, intent, outcomes, updates and risks are to be considered basic.

• Human alternatives, consideration and fallback: In short, when a user determines that an automated system has become too unwieldy or its functionality too untenable, then he or she should be able to have access to a real person to help them. No one should feel trapped within the confines of an all-powerful system they do not understand and cannot properly operate.

These principles could become a friendly conversation starter. As citizens, we need a simple tool to unify the discussion as we confront this significant challenge. This AI Bill of Rights could be it.

These articles are being shared by partners in The Granite State News Collaborative. For more information visit collaborativenh.org.

]]>