Getting AI experts and economics experts together to figure out the economic impacts of transformative AI (TAI).
getty
In today’s column, I examine a newly released research paper that tackles an important topic, namely, the need to formulate and promulgate a big picture perspective regarding the economic and societal impacts of transformative AI.
The paper was recently posted by the esteemed National Bureau of Economic Research (NBER) and does a yeoman’s job in laying out an engaging and foundational big picture or framework that deserves keen consideration. I will walk you through the key aspects and aim to whet your appetite on the altogether weighty matter.
We definitely need more work of this kind. The economic upheaval that might very well coincide with the rise of artificial general intelligence (AGI) and someday artificial superintelligence (ASI) requires rapt attention now. We can’t put off these crucial analyses. The usual refrain by high-tech is that we should mindlessly move fast and break things. But misguidedly breaking our economies and economic formations carries enormously adverse consequences, especially if we aren’t preparing ourselves for the consequences.
Let’s talk about it.
This analysis of AI breakthroughs is part of my ongoing Forbes column coverage on the latest in AI, including identifying and explaining various impactful AI complexities (see the link here).
Jumble Of Economic Concerns
You are likely bombarded daily with news encompassing dire economic handwringing when it comes to the future of AI.
Headlines declare that most jobs will be utterly replaced by AI. Will your job be on the cutting block? Is your career path coming to a dead end? What should those newbies who are newly entering the workforce be doing to try and get a leg up on the work-related Armageddon arising from AI adoption?
Some have been declaring that we will need to institute a form of universal basic income (UBI). This means that everyone will get paid something by the government, partially to offset the massive layoffs that are being predicted. Will those who are working only need to work for 2 or 3 days a week? Maybe none of us will work at all.
On the upside, maybe AI will be a grandiose inventor and discover new inventions that will bolster humanity immensely. AI might find cures for cancer. AI might devise practical ways to ensure that everyone gets a workable and affordable jetpack or possibly create ready-to-go flying cars. For my watchful rundown of some of the anticipated larger-than-life inventions that advanced AI might invent, see my analysis at the link here.
Will the upsides, such as AI inventing incredible devices and cures, possibly impact our economies and economic conditions?
Absolutely.
You see, AI as dual-use technology is going to simultaneously aid us and undercut us. The upsides have economic impacts, and likewise, the downsides do too. There are monumental economic considerations afoot. Those who take a conventional tech-only viewpoint tend to downplay or even ignore the emerging ramifications of AI. A head-in-the-sand approach to economics and societal impacts associated with advanced AI is a surefire recipe for disaster.
We must bring the economists into the fold and get their insights into the future of AI, which ties to the future of humankind, and figure out where we are, where we are headed, and what ought to be done to steer us toward a bright economic future rather than a dismal or collapsing one.
Need The Big Picture Framework
Economic considerations about AI are scattered here and there. Often, a particular economic factor is selectively assessed, such as questions about jobs, but this is not done in concert with other related factors. It is akin to picking up one piece of a puzzle and failing to ascertain how the piece fits into the larger assembly.
Jose Saramago, winner of the Nobel Prize in Literature, notably made this remark about how to respond to situations that are jumbled: “Chaos is merely order waiting to be deciphered.”
I will be addressing two realms of structure or deciphering associated with the impacts of advanced AI:
- My impact levels that are associated with AI transformations.
- Recent research paper offering a vital framework for economic analysis regarding AI transformations.
Let’s jump in.
Impact Of Large-Scale AI Transformations
There is a great deal of disagreement about the likely magnitude of the economic and societal impacts of advanced AI.
We shall start with some overall public opinion. Most laypeople would probably say that advanced AI, especially pinnacle AI such as AGI, is going to have substantive impacts on their lives. For my detailed indication of what AGI will consist of, see the link here, and for my year-by-year prediction of when AGI will be attained, see the link here.
Not everyone necessarily believes that the magnitude is going to be of a pronounced nature. Some would concede that there will be a bit of an impact, though in the same breath, they would argue that we don’t need to be overly concerned. AI will end up being a minimal form of economic transformation. No worries.
Others would have their eyes popping out of their head at such a laissez-faire decree. To say that pinnacle AI is going to be minimally impactful seems completely out-to-lunch. AI is going to have a rippling effect. It will be like tossing a large boulder into a small pond. We need to anticipate where the waves are going to crash.
At the farthest end of this spectrum are those who believe AI will be extraordinarily transformative. Existing economies will be totally reshaped in radical ways. The new norm is going to be far beyond what we have now. Everyone needs to look at the future with the realization that humanity will be faced with incredible challenges that have never been encountered in the entire history of humankind.
Defining AI Impact Levels
How can we reconcile these varying viewpoints?
First and foremost, the underlying viewpoint must be made visible and explicit.
When I give various talks about trends in AI, I often share my set of AI impact levels, doing so to set the stage for sensibly discussing the future. This is reminiscent of talking about autonomous systems such as self-driving cars. You might be aware that self-driving cars are ranked based on their levels of autonomy.
In terms of AI impact levels associated with the transformational criteria, here are my devised dimensions that range from 0 (low) to 3 (high):
- AI Impact Level 0: Not transformational. Advanced AI will have a negligible or near-zero impact.
- AI Impact Level 1: Minimally transformational. AI will have a “normal” tech-driven impact (for the ongoing debate about whether AI should be construed as so-called normal tech, see my analysis at the link here).
- AI Impact Level 2: Semi-transformational. AI will have an above-normal semi-transformational impact, including pronounced rippling impacts.
- AI Impact Level 3: Full-on transformational. Transformations will be severe, widespread, and will deeply change the nature of our economic and societal conditions.
Where do you stand?
Few people would pick Level 0 since it seems nearly obvious that advanced AI is going to rise above a floor-level of being impactful. The Level 0 camp is either unwilling to entertain rational discussions or has opted to be dogmatic and unyielding in their chosen posture.
We can somewhat gingerly set aside Level 0 and perhaps generally agree that we are facing either Level 1, Level 2, or Level 3. Frankly, I reject the Level 1 argument as stated in my analysis referenced above and lean toward Level 2 or Level 3. All in all, my best guess is that Level 3 is where we are headed.
Handy Framework On The Economics
In the next step of turning chaos into something orderly, we need to have a cogent framework that addresses the economic and societal impacts of advanced AI. To refer to AI when it has a transformative likelihood, we shall use the moniker of TAI, i.e., transformative AI.
The research paper that contains such a framework and that I earlier alluded to is entitled “A Research Agenda for the Economics of Transformative AI” by Erik Brynjolfsson, Anton Korinek, and Ajay K. Agrawal, National Bureau of Economic Research (NBER) Working Paper Series, September 2025, and made these salient points (excerpts):
- “As we approach Transformative Artificial Intelligence (TAI), there is an urgent need to advance our understanding of how it could reshape our economic models, institutions, and policies.”
- “Economics provides powerful frameworks for analyzing how transformative technologies affect society through production processes, resource allocation, and market mechanisms.”
- “Economists study how technologies like AI change production functions — the relationship between inputs (labor, capital, data) and outputs — and how these changes ripple through markets, affecting productivity, wages, prices, and growth.”
- “We present a research agenda to motivate attention to critical questions about how TAI will impact the economy and societal well-being.”
- “Our goal is to inspire the research community to tackle these challenges and questions.”
The research paper is a great read and worth digging into. I will provide selected highlights to pique your interest.
Being Tangible And To The Point
First, to set the stage, one aspect that gets on my nerves consists of economic analyses that wiggle around and won’t commit to anything tangible. Those kinds of write-ups are frustrating since you cannot hold them accountable for their various proclamations. It is all mushy.
An element of this newly released paper that I especially favored is the attention to being specific and tangible. For example, consider this definition of a high-level measure or metric for determining what TAI entails (excerpts):
- “From an economic perspective, we define Transformative AI as artificial intelligence that enables a sustained increase in total factor productivity growth of at least 3 – 5x historical averages.”
- “Total factor productivity growth measures the rate at which an economy’s output increases beyond what can be explained by additional inputs of labor and capital.”
- “Such growth may occur because AI facilitates a radical new set of goods, services, or production processes, because AI changes the relative scarcity of inputs, particularly by making cognitive labor significantly more abundant relative to other factors, or because AI creates novel economic organizations and institutions.”
You can plainly see that there is no place to hide since the paper tells us straightforwardly the economic measurement for TAI. Boom, drop the mic.
Of course, some might not concur with that definition and opt to debate whether it is the right measure. I’m good with that. It is better to have placed a stake in the ground and allow others to agree or disagree, versus being vacuous and unspecific.
I prefer bravery in these evolving matters.
Set Of Grand Challenges
Let’s keep going on the heralded path of being specific.
The research paper posits that there are nine grand challenges associated with TAI:
- Economic Growth
- Invention, Discovery, and Innovation
- Income Distribution
- Concentration of Decision-making and Power
- Geoeconomics
- Information, Communication, and Knowledge
- AI Safety & Alignment
- Meaning and Well-being
- Transition Dynamics
For each of those grand challenges, the paper presents essential questions that need to be explored.
I’ve been covering many of those challenges in my column. For example, if you are interested in AI safeguards and human-values alignment, see my discussion at the link here and the link here, just to name a few.
Once again, a smarmy person might exhort that there are more than nine such grand challenges. Or they might insist that subordinated topics in some of the nine should be placed on their own accord as a grand challenge, such as environmental sustainability, cultural pluralism, weaponization, and so on.
The gist is that we have a strawman and can converse intelligently on something palpable, rather than talking in circles due to vagaries and non-specifics.
Decisive Approaches To TAI
The paper provides a set of six approaches that might be pursued to further deepen and extend the TAI economics perspective. Researchers interested in picking up the ball and running with it can potentially select one or more of the stipulated approaches.
In my own words, the elucidated approaches generally consist of:
- Utilize economic theories and adapt them to TAI as needed
- Devise and adopt a TAI Economic Tracking dashboard
- Consider crafting new metrics for assessing TAI economic impacts
- Ensure that both macroeconomic and microeconomic perspectives are brought into the picture.
- Perform economic-oriented simulations about TAI by using AI
- Undertake expert-informed scenario planning and TAI futures forecasting
AI researchers who want in on this action would probably find that #2 on devising a tracking dashboard might be up their alley, and the same could be said about #5, involving crafting AI-based simulations covering these matters. That being said, the best bet is to have both AI researchers and economics researchers putting their heads together in a multi-disciplinary way. It takes two to tango, and each side needs the other.
Behind The Eight Ball
What happens when you get AI experts and economics experts in the same room?
I would hope that they would work smartly side-by-side on how we are going to cope with pinnacle AI. There is no time to waste. Even if advanced AI is slower to materialize than we assume, figuring out and preparing for the economic consequences is a big task. In addition, there are actions we can take now that will shape where the future is going to land.
As Mark Twain eloquently stated: “The secret of getting ahead is getting started.” Research papers of this nature are a welcome sign that we are indeed getting solidly underway. Let’s keep the momentum going.
It’s for the sake of humankind that we must do so.