We need to address the ‘pacing problem’ before AI gets out of control
Frazier is an assistant professor at the Crump College of Law at St. Thomas University. He previously clerked for the Montana Supreme Court.
The "pacing problem" is the most worrying phenomenon you've never heard of but already understand. In short, it refers to technological advances outpacing laws and regulations. It's as easy to observe as a streaker at a football game.
Here's a quick summary: It took 30 years from the introduction of electricity for 10 percent of households to be able to turn on the lights; 25 years for the same percentage of Americans to be able to pick up the phone; about five years for the internet to hit that mark; and, seemingly, about five weeks for ChatGPT to spread around the world.
Ask any high schooler and they’ll tell you that a longer deadline will lead to a better grade. Well, what’s true of juniors and seniors is true of senators and House members – they can develop better policies when they have more time to respond to an emerging technology. The pacing problem, though, robs our elected officials of the time to ponder how best to regulate something like artificial intelligence: As the rate of adoption increases, the window for action shrinks.
A little more than a year out from the release of ChatGPT, it’s already clear that generative AI tools have become entrenched in society. Lawyers are attempting to use it. Students are hoping to rely on it. And, of course, businesses are successfully exploiting it to increase their bottom lines. As a result, any attempt by Congress to regulate AI will be greeted by an ever expanding and well-paid army of advocates who want to make sure AI is only regulated in a way that doesn’t inhibit their client’s use of the novel technology.
ChatGPT is the beginning of the Age of AI. Another wave of transformational technologies is inevitable. What’s uncertain is whether we will recognize the need for some regulatory imagination. If we stick with the status quo – governance by a Congress operated by expert fundraisers more so than expert policymakers – then the pacing problem will only get worse. If we instead opt to use our regulatory imaginations, then there’s a chance that future surges in technology can be directed to align with the public interest.
Regulatory imagination is like a pink pony – theoretically, easy to spot; in reality, difficult to create. The first step is to encourage our regulators to dream big. One small step toward that goal: Create an innovation team within each agency. These teams would have a mandate to study how the sausage is made and analyze and share ways to make that process faster, smarter and more responsive to changes in technology.
The second step would be to embrace experimentation. Congress currently operates like someone trying to break the home run record – they only take big swings and they commonly miss. A wiser strategy would be to bunt and see if we can get any runners in scoring position; in other words, Congress should lean into testing novel policy ideas by passing laws with sunset clauses. Laws with expiration dates would increase Congress’ willingness to test new ideas and monitor their effectiveness.
Third, and finally, Congress should work more closely with the leading developers of emerging technologies. Case in point, Americans would benefit from AI labs like OpenAI and Google being more transparent with Congress about what technology they plan to release and when. Surprise announcements may please stakeholders but companies should instead aim to minimize their odds of disrupting society. This sort of information sharing, even if not made public, could go a long way toward closing the pacing problem.
Technological “progress” does not always move society forward. We’ve got to address the pacing problem if advances in technology are going to serve the common good.
- How artificial intelligence can be used to reduce polarization ›
- Why are Black women becoming the hidden figures in AI? ›
- Put AI on the ballot ›
- The critical need for transparency and regulation amidst the rise of powerful artificial intelligence models ›