🤖 What's really at stake if we get AI regulation wrong

After a half-century of economic growth far slower than what it could have been AND should have been, we don’t need more needless delay.

Quote of the Issue

“An Up Wing future depends on us getting the most possible out of our cities. It always has. Urban economist Edward Glaeser has explained that these ‘dense agglomerations that dot the globe have been engines of innovation since Plato and Socrates bickered in an Athenian marketplace.’ And to study them “is to study nothing less than human progress.” - James Pethokoukis, The Conservative Futurist: How to Create the Sci-Fi World We Were Promised

---

Some shameless self-promotion: I have a new book out. The Conservative Futurist: How To Create the Sci-Fi World We Were Promised is currently available pretty much everywhere. I’m very excited about it! Let’s gooooo! 🆙↗⤴📈

• Amazon

• Barnes & Noble

• Books-A-Million

• Target

• Walmart

• Bookshop

---

The Essay

🤖 What's really at stake if we get AI regulation wrong

Item: The Biden administration plans to invoke emergency federal powers as part of a new executive order aimed at reining in the risks of artificial intelligence, a new technology as powerful as it is potentially disruptive. President Biden will release on Monday an order invoking the Korean War-era Defense Production Act which would compel major AI companies to notify the government when developing any system that poses a “serious risk to national security, national economic security or national public health and safety,” according to fact sheet that White House aides shared over the weekend. The order aims to step into a global regulatory vacuum over a fast-growing technology. It seeks to mitigate risks from privacy to job losses and underscores an effort to address the sometimes slow pace of response to the new technology among those in Congress and other government agencies. - “Biden to Use Emergency Powers to Mitigate AI Risks,” The Wall Street Journal, 10/30/2023

Here are four of my baseline assumptions regarding artificial intelligence/machine learning/generative AI:

• First: GenAI is potentially a powerful general-purpose technology, one as important as the Information and Communication Technology Revolution (PC + internet combo) — but also perhaps far more important and transformative. (More on that below). It could decisively end what some economists call the Great Stagnation, and what I call the Great Downshift. (Time for the Great Upshift!) Every other country wishes it had an AI sector as advanced as America’s.

• Second: We might well not yet have GenAI — or a number of other emerging technologies in biotech, energy, and space — if not for the ICT Revolution making it easier for scientists and technologist to find and access information and then share it with colleagues. Communication + collaboration.

• Third: Key to the ICT revolution was our decision in the 1990s to adopt a hands-off approach to the internet. As The New York Times reported back in July 1997:

  Under pressure to quickly build a secure online marketplace for American business, President Clinton pushed Tuesday to make the Internet a "global free-trade zone" and promised to put key patent, privacy and copyright policies in place within 12 months. He and Vice President Al Gore were hosts for an East Room reception of industry executives to endorse a Presidential task force's call for a hands-off, no-new-taxes approach to regulating business transactions on the worldwide computer network. The Internet "should be a place where government makes every effort . . . not to stand in the way, to do no harm," Clinton said.

• Fourth: Due to an obsessive focus on the downsides of social media platforms and a deep techno-pessimist streak in our culture for the past half-century, there’s a dangerously high probability that government is overly aggressive when it comes to AI regulation.

So given the above, it’s not surprising that I’m concerned about President Biden’s new executive order on AI, about which White House aide Bruce Reed said in a statement: "Biden is rolling out the strongest” set of actions any government in the world has ever taken on AI safety, security and trust. It's the next step in an aggressive strategy to do everything on all fronts to harness the benefits of AI and mitigate the risks."

Regulatory escalation

Does the content of the order match the Biden White House’s framing? Well, it’s certainly a massive escalation in regulatory aggressiveness. For example: The order moves beyond the “previous voluntary commitments that the White House garnered from AI companies and requires notification in accordance with the Defense Production Act,” according to Axios. Also some great points by technology policy analyst Adam Thierer in a new blog post, in which he argues that the Biden order seems to allow federal agencies to subtly transform voluntary AI guidelines into a more covert regulatory framework, especially given Congress's inaction on AI matters. Furthermore, the order's endorsement for the FTC to intensify its focus on AI policy raises concerns about potential overreach, given the FTC's expansive regulatory powers and current eagerness to use them.

My view: Premature and rushed AI regulation risks stifling innovation and cementing dominant companies, especially as the major players have the resources and clout to deal with new rules and to influence the shaping of those rules to their advantage. Regulators should show GREAT humility, given the limited understanding of generative AI's risks.

Sure, governments should prioritize establishing structures to study AI, encouraging collaboration among existing regulatory bodies. Voluntary codes of conduct for AI model-makers can help manage potential threats. And then we can take it from there, dealing with problems in an informed and targeted way. There’s no reason not to view the 1990s internet example as informative.

“The development of capable AI is likely to be among the most consequential macroeconomic stories of the 21st century, with important implications for relative economic performance, financial market returns, and longer-run interest rates.” - Goldman Sachs

Thinking about the risks from getting AI regulation wrong

Let’s remember what’s at stake here: a technology that could accelerate technological progress and economic growth, broadly boosting incomes, health, and societal resilience. As Goldman Sachs economists declare in “Upgrading Our Longer-Run Global Growth Forecasts to Reflect the Impact of Generative AI,” a brand-new research note from the bank: “In our view, the development of capable AI is likely to be among the most consequential macroeconomic stories of the 21st century, with important implications for relative economic performance, financial market returns, and longer-run interest rates.”

That said, GS is actually pretty cautious, as I see it. Its initial analysis suggests that the use of advanced AI could boost the global economy by 10-15 percent over the long run. In an earlier report, the bank offered a baseline forecast that AI could contribute 1.5 percentage points to annual US productivity growth if widely adopted over 10 years.

But there are some headwinds identified in this report. AI might merely replace some of the economic growth currently being generated by existing info and communications technologies. (GS estimates that “almost half of all labor productivity growth over the last 20 years can, on average, be attributed to advancements and investment in ICT.)