Is It Time to Regulate AI?
Artificial intelligence is fast becoming a part of being both a consumer and an employee. Apply for a credit card or mortgage and many banks will use AI to weigh your creditworthiness. Apply for a job and the employer could use AI to rank your application. Call a customer-service number and AI technology might screen and route your call.
AI refers to techniques that allow computers to learn, reason, infer, communicate and make decisions that in the past were the sole province of humans.
Yet as AI technology spreads, so do concerns about its accuracy and fairness. Experts say it can have built-in racial, gender and age biases that could, for instance, rule out certain qualified people for jobs, or force some creditworthy borrowers to pay higher rates than otherwise. This has prompted calls for regulation, or at least greater transparency about how the systems work and their shortcomings.
We asked three experts to discuss the issue: Esha Bhandari, deputy director of the American Civil Liberties Union’s Speech, Privacy and Technology Project; Ryan Calo, a professor at the University of Washington School of Law; and Jordan Crenshaw, vice president of the U.S. Chamber of Commerce’s Technology Engagement Center.
What follows are edited excerpts of the discussion, which took place over email.
WSJ:Do we need greater regulation of artificial intelligence? Or just government guidance that would lead the industry to regulate itself?
MS. BHANDARI: Government guidance has a role to play, but where business use of AI has the potential to harm individuals or communities, those harms have to be addressed through regulation. It’s the same principle that applies to other consumer products that have the potential to harm people—businesses that stand to make money from those products are not left to simply self-regulate.
One argument that business often uses to avoid regulation is that if the public is unhappy with their product or service, the public will take its business elsewhere. That’s often not possible with AI tools that are developed by companies that don’t sell to the public and are impervious to those pressures.
MR. CRENSHAW: Artificial intelligence has the potential to significantly improve society, like securing our networks, preventing fraud, expanding financial inclusion and helping medical researchers develop treatments quicker. At the same time, every iteration of technology has risks.
If we rush to outright ban or overregulate AI, we will delay or won’t realize these new societally beneficial uses. There may come a time where we find we need to regulate AI—as is the case with privacy—but we should really thoughtfully approach this issue before rushing to regulate.
This is one of the reasons why the U.S. Chamber of Commerce launched its new AI Commission on Competitiveness, Inclusion and Innovation, to study how best to approach AI from a regulatory perspective by hearing directly from all relevant stakeholders, including consumer advocates, business and academia. This commission will release a report this fall.
MS. BHANDARI: From a civil liberties and civil rights perspective, I would disagree that we need to develop and deploy AI first and ask questions later. In the medical or pharmaceutical context, we have strong reasons to spur advancement while acknowledging that new treatments or drugs cannot simply be unleashed on the public without satisfying regulatory standards first.
MR. CALO: I’m not convinced it’s possible to regulate artificial intelligence as such. AI isn’t a thing, like a train, but rather a set of techniques aimed at approximating some aspect of cognition. But that doesn’t mean there shouldn’t be changes to law. The ability of AI to spot patterns in peoples’ data, for example, suggests a need for tougher privacy laws. The disparate impact AI can have on marginalized consumers or job seekers suggests a role for federal agencies to address bias.
Ultimately, proponents of AI can’t have it both ways.
