The Senate Judiciary Committee recently held a hearing on regulating Artificial Intelligence, a topic that has gained significant traction in Washington with the rise of ChatGPT. As is often the case with new technologies, fear of the unknown is driving a familiar pattern of scaremongering followed by overregulation. The mother-may-I mavens are moving rapidly to stifle AI before we fully understand its applications and risks.
The nearly three-hour panel was dominated by Samuel Altman, CEO of the nonprofit OpenAI that released ChatGPT to the public earlier this year. There was some productive discussion of intellectual property and privacy protections, both of which are real concerns that Congress should address. Unfortunately, lawmakers mainly focused on proposing new ways to regulate AI during the hearing, while next to nothing was said about new opportunities or economic development.
Altman’s chief suggestion was for the government to “license” future AI developers. This would mean that after OpenAI has already developed cutting-edge technology, anyone else who wants to build their own version needs permission from a faceless government bureaucrat. It is not a surprise that Altman, looking to strangle potential competitors in the crib, would pull the ladder up behind him in this fashion.
Instead of using new innovations as justification to expand the size and scope of government, lawmakers should instead keep several first principles in mind while building a consistent, light-touch framework to regulate AI.
Regulatory Humility is Necessary to Beat Communist China in AI Race
The genie is not going into the bottle – the United States is not the only country developing AI. In 2017, the Chinese Communist Party announced that it intends to be the global leader in AI by 2030. Recent estimates indicate that this could add over $600 billion to the Chinese economy annually.
While the full benefits of AI remain unclear, one fact remains true – every past leap in technology has yielded more benefit than harm in the long run. Some initial opportunities evident today include AI’s potential to predict natural disasters in areas that flood frequently, better statistical analysis to diagnose cancer, and more helpful smartphones. Analysts project that AI will increase global GDP by 7 percent this decade, including a 1 percent increase in the US alone (more than $2 trillion).
The real gains will come from innovations no one can currently envision. Imagine traveling back to 1776 and trying to explain to the British Parliament that James Watt’s new steam engine for coal mining will eventually lead to electricity, family cars, and suburbs. That’s where we are now, and regulatory humility is necessary to fully unleash the benefits of AI that remain unclear today.
Even if we wanted to freeze technology in the year 2021, the Chinese Communist Party will have no such compunctions and will gladly smash through all our cybersecurity measures if we stop innovation. AI is here to stay whether we like it or not, so the question is how to ensure neutral principles like intellectual property, privacy, and civil liberties are honored in this new world.
Protecting Intellecutal Property Remains Paramount
IP protections are inherently good and will help enable many of the future advances in AI. Policymakers should embrace innovation with light guardrails that encourages creativity. Our country was built on permissionless innovation and the same principle should prevail here, provided that “advancements in the useful arts” (as the U.S. Constitution calls them) are justly rewarded through IP protections.
Exactly how to do this is not yet clear, but one possibility is blockchain and future iterations of blockchain-like technologies. These can provide unique tokens to prove ownership of digital assets, and not just pictures of apathetic apes. Provided the same logic for property rights is applied, non-intrusive guardrails can be maintained. This will incentivize more and better innovations in the next few years.
Congress Should Continue Work on National Privacy Law
On the issue of privacy, AI is only the latest wrinkle in the long-running debate about a national framework. Users agreeing to trade their personal data for use of tech platforms is a cornerstone of the modern economy, and data monetization was worth more than $2 billion in 2022. While estimates vary, it is forecast to surpass $9 billion by 2030. Users freely agree to this arrangement, but exactly how their personal information is handled, stored, and traded by AI is unknowable at this point.
Rather than stifle development, however, Congress should see the rise of AI as another reason to finally pass a national data privacy law. Setting one national standard, rather than a patchwork of state standards, will benefit most domestic industries (bankers, for instance, have been asking Congress to normalize standards for years) and clearly falls under purview of interstate commerce per the Constitution.
If Congress continues to delay action, a 50-state privacy patchwork could cost $1 trillion over 10 years, with $200 billion falling on small businesses. Now is a better time than ever for Congress to work out a reasonable, consistent standard that enables economic growth while protecting sensitive information.
Civil Liberties Must Be Preserved
In what may be the gravest danger that AI poses, government bureaucrats could soon have unprecedented access to our personal effects. We have already seen what can happen when federal law enforcement agencies pressure social media companies censor Americans’ speech, and it seems we are subjected to annual revelations of intrusive data collection by national security agencies. Even when telephones were relatively young, the FBI needed strict oversight of its phone-tapping operations. In the wrong hands, AI will only enhance surveillance and censorship powers.
Fortunately, we already have a blueprint for protecting Americans’ rights to freedom of expression, due process, and private property: the U.S. Constitution. In the in the 235 years since the Constitution was ratified, these enumerated protections have been buttressed by a large body of legislation and judicial precedent to shield the citizenry from the abuses of the state. Once again, Congress must adapt our laws to meet the moment and ensure that algorithms respect our civil liberties.
Workplaces and classrooms will have to rapidly adapt to the challenges that will inevitably result from machines impressively mimicking human behavior. There are wrinkles to be ironed out. But permissionless innovation has long been the key to American economic dynamism, and it must continue with AI.
While Altman and his confederates advocate a “pause” on AI development, others have suggested we should do the reverse: we should “pause” politicians from regulating AI for a period of several months, with opportunity for renewal upon expiry. Once this cooling-off period is over, perhaps we can have a serious discussion of what guardrails will best protect consumers while stimulating innovation and growth.