TorontoStartsBook a call

Anthropic’s AI Economy Model: 2030 Jobs, Wages, GDP and Market Risks

Silicon Scoop · Episode 571 · November 4, 2026

Anthropic AI economy model forecasts 2030 GDP, jobs and wages, while AI unemployment projections and labor-share risks challenge policymakers.

With Craig Major and Jon Irwin

Open audio in your browser
Anthropic’s AI Economy Model: 2030 Jobs, Wages, GDP and Market Risks

Anthropic's AI Economy Model Just Mapped 2030 Jobs and Wages

Estimated reading time: 10 minutes

Key Takeaways

  • The Anthropic AI economy model, formally titled Economic Scenarios for Transformative AI (Anthropic Institute Working Paper No. 2026-02), is a structured set of scenarios — a planning tool — built to help policymakers, economists, and business leaders understand what different levels of AI capability and adoption could mean for the U.S. economy over the next four years.
  • The research team built three core scenarios: Modest AI Growth, Substantial AI Growth, and Extreme AI Growth.
  • In the extreme scenario, cognitive worker unemployment reaches 17.9%, cognitive wages fall 11.5% below the no-AI path, and labor's share of income collapses from 60% to 45.2%.
  • The model explicitly omits policy responses, business cycles and aggregate-demand effects, financial market disruption, data center buildout effects on the broader economy, rapid robotics progress, and catastrophic AI risks or an AI investment bubble.

Table of contents

The numbers coming out of Anthropic's new economic model are the kind that stop a room cold.

Not because they predict doom. Not because they predict a golden age. But because they lay out, in cold mathematical detail, exactly how much AI could reshape who earns money, how much they earn, and whether they have a job at all by 2030.

The Anthropic AI economy model, formally titled Economic Scenarios for Transformative AI (Anthropic Institute Working Paper No. 2026-02), is not a crystal ball. The researchers are clear about that. It is a structured set of scenarios — a planning tool — built to help policymakers, economists, and business leaders understand what different levels of AI capability and adoption could mean for the U.S. economy over the next four years.

What it found should be required reading for anyone who works with their brain for a living.

What the Anthropic AI Economy Model Actually Is — and What It Is Not

Before we get into the numbers, context matters enormously here.

This is not a forecast. It is not a probability-weighted prediction. The Anthropic team, led by economists Anton Korinek, Charles I. Jones, Szymon Sacher, Tess Cotter, and Peter McCrory, built a framework of scenarios ranging from modest to extreme. Each scenario asks a simple question: "If AI capability and adoption reached this level by 2030, what would the economy look like?"

The model is calibrated to a mid-2026 baseline. At that starting point, AI affects roughly 14% of economy-wide tasks and is actively used in about 10% of instances within those affected tasks. The pre-AI economic benchmark is anchored to 2024 data, and the model simulates the economy on a monthly basis through 2030.

The technical backbone of this model is serious work. It uses the U.S. Department of Labor's O*NET occupation-task taxonomy to divide the workforce into two broad categories:

  • Cognitive workers: management, professional, sales, and office roles
  • All other workers: jobs less exposed to AI automation

It treats every job as a bundle of tasks. Each task can be augmented by AI, fully automated by AI, left unchanged, or replaced by entirely new tasks that AI creates. The model uses a task-level CES (Constant Elasticity of Substitution) production function with an elasticity of substitution of 0.5, a baseline labor-income share of 60%, an elasticity of capital supply of 3, and a monthly search-and-matching labor-market block.

These aren't just academic details. These assumptions are the levers that determine how productivity gains get split between higher wages, better capital returns, and unemployment. Change the assumptions, and the outputs shift dramatically — which is exactly the point of the scenario framework.

Critically, the model explicitly omits:

  • Policy responses from governments
  • Business cycles and aggregate-demand effects
  • Financial market disruption
  • Data center buildout effects on the broader economy
  • Rapid robotics progress
  • Catastrophic AI risks or an AI investment bubble

This means the model cannot tell you what happens if the AI hype cycle collapses, or if a recession follows mass layoffs, or if AI does something truly catastrophic. Reviewers including economists Daron Acemoglu, David Autor, Emi Nakamura, and David Romer flagged these limitations directly — including coarse worker heterogeneity, uncertain occupational shrinkage, and missing aggregate-demand effects from data center investment.

Sources: Anthropic Institute econ-scenarios | Anthropic working paper PDF

The AI Jobs Forecast 2030 — Three Scenarios, Three Very Different Americas

This is where the model gets genuinely important. The research team built three core scenarios. They diverge mainly after 2027, because all of them start from the same mid-2026 readings and only separate as capability, adoption, productivity, and automation assumptions are dialed up.

Scenario 1: Modest AI Growth

In this scenario, AI affects 20% of cognitive-task mass and reaches 20% diffusion across the economy.

The results are nearly invisible at the macro level:

  • GDP ends up 1.6% above the no-AI path
  • Unemployment nudges from 3.8% to 3.9%
  • Wages remain broadly stable

This is the version of the future where AI is a useful productivity tool — like spreadsheets or email — that makes workers somewhat more efficient without dramatically altering who does what or who earns what. Most people's day-to-day economic experience changes very little.

Scenario 2: Substantial AI Growth

In this scenario, AI affects 30% of cognitive-task mass and reaches 40% diffusion, automating 75% of AI-performed task instances.

Now the numbers start to move:

  • GDP rises 8.3% above the no-AI path
  • Overall unemployment reaches 4.6%
  • Cognitive worker wages fall by 0.3%
  • Labor's share of national income drops from 60% to 56.1%

This is a meaningful shift. A near-$1 trillion-scale GDP gain accompanied by wage pressure on knowledge workers and a growing share of income flowing to capital owners rather than workers. It is not a catastrophe, but it is a redistribution — quiet, gradual, and easy to miss until it has already happened.

Notably, Anthropic surveyed 10,980 U.S. adults through Morning Consult between August 11–23, 2026. The median respondent's answers implied an outcome close to this substantial scenario: GDP 8.6% above the no-AI path and overall unemployment around 4.6%. Ordinary Americans, in other words, are expecting the middle path.

Scenario 3: Extreme AI Growth

This is the thought experiment that has been making headlines. AI affects 50% of cognitive-task mass, reaches 60% diffusion, and automates 90% of affected task instances.

The outputs are staggering:

  • GDP rises 32.4% above the no-AI path
  • Annual GDP growth hits 15.4% — historically unprecedented
  • Overall unemployment reaches 11.9%
  • Cognitive worker unemployment reaches 17.9%
  • Cognitive wages fall 11.5% below the no-AI path
  • Labor's share of income collapses from 60% to 45.2%
  • The capital stock sits 56.3% above the no-AI path
  • Capital income is 81.4% higher

In plain terms: the economy becomes dramatically larger, but almost all of the new wealth flows to the owners of capital — the companies, the data centers, the compute infrastructure — rather than to the people doing the work. Knowledge workers, the very group that has been insulated from automation for decades, face nearly one-in-five unemployment.

The AI labor share question at the heart of this scenario is stark. If GDP grows by a third but wages fall and unemployment nearly triples, the aggregate statistic of "economic growth" masks an enormous transfer of income from workers to capital holders.

Sources: Anthropic working paper PDF | Anthropic Institute econ-scenarios

Stop Running Your Business on Manual Work

The economic shifts ahead — whether modest or extreme — mean one thing is certain: businesses that still rely on repetitive manual processes will fall further behind, faster than ever. While companies debate AI's macro effects, smart operators are already automating their workflows and reclaiming hours every week.

Flowgrammer helps Canadian businesses automate the tasks that eat your team's time — so you can scale without adding headcount.

👉 See What You Can Automate Today

AI Unemployment Projections and What They Mean for Knowledge Workers Specifically

The AI unemployment projections in the substantial and extreme scenarios carry a specific warning that deserves its own conversation: this automation wave is aimed squarely at cognitive workers.

This is a reversal of historical patterns.

Previous waves of automation — from the loom to the assembly line to early computing — primarily displaced physical and routine clerical labor. White-collar professionals, managers, analysts, writers, lawyers, and programmers were largely protected because their work required judgment, creativity, and complex communication. Machines could not do those things.

Generative AI can now do versions of many of those things. The O*NET task taxonomy that Anthropic used to build this model identifies management, professional, sales, and office occupations as the primary zone of exposure. These are, by definition, the jobs that traditionally commanded higher wages and offered more economic security.

In the extreme scenario, cognitive worker unemployment of 17.9% represents an almost inconceivable labor market disruption for a group that has historically enjoyed unemployment rates well below the national average.

What drives this? Three compounding factors in the model:

  1. Task automation rate: AI does not just assist — it fully replaces the task, eliminating the need for the worker to perform it at all
  2. Diffusion speed: The faster AI spreads across industries, the less time workers have to retrain, relocate, or move into new roles
  3. Low elasticity of substitution: The model's elasticity of 0.5 means that when AI becomes very cheap relative to labor, firms substitute heavily toward capital and AI, rather than sharing productivity gains with workers through higher wages

Anthropic economist Anton Korinek notes that AI capability may advance rapidly but economic diffusion could be slower — meaning the extreme scenario may be technically possible but economically gradual in practice. Co-founder Jack Clark has separately argued that policymakers should prepare spending mechanisms in advance, because unusually high growth would create substantial fiscal capacity that could fund retraining, transfers, and social insurance programs.

Researcher Peter McCrory makes perhaps the most important point of all: these outcomes are not predetermined. Society retains real agency through adoption choices, regulatory decisions, and deliberate policy. The model shows what could happen given specific assumptions — not what will happen if humans make active choices.

The practical implication for professionals and businesses right now is clear: the AI jobs forecast 2030 points toward a labor market that rewards workers who can work alongside AI rather than compete against it — and organizations that build AI-augmented workflows now will have a significant head start.

Sources: Anthropic Institute econ-scenarios | NPR coverage | Axios coverage

What the Model Leaves Out — and Why That Matters as Much as What It Includes

The explicit omissions in the Anthropic model are not a flaw. They are an honest acknowledgment of complexity. But understanding them is essential to using the model responsibly.

The missing pieces include:

Policy responses. The model assumes no government intervention — no retraining programs, no universal basic income, no changes to tax policy, no new labor regulations. Real economies do not work this way. Policy will respond to disruption, and the character of that response will heavily shape whether the substantial or extreme scenario results in broadly shared growth or a concentrated wealth shock.

Aggregate-demand effects. If millions of knowledge workers lose income, they spend less. That spending reduction feeds back into the broader economy, potentially triggering recessions that the model's GDP figures do not capture. Reviewers Acemoglu and Autor flagged this gap specifically.

Data center buildout effects. The extreme scenario requires an enormous acceleration of compute infrastructure. The economic activity generated by building that infrastructure — the jobs, the supply chains, the energy demand — is not modeled. It could partially offset some of the wage effects.

An AI investment bubble. The model cannot simulate what happens if the current wave of AI investment significantly exceeds actual productivity returns — a scenario that has precedent in the dot-com era. A sharp correction in AI valuations could produce economic damage entirely separate from the labor displacement effects modeled here.

Catastrophic AI risks. Notably, Anthropic's own alignment science lead, Evan Hubinger, has publicly estimated a greater than 10% chance of AI causing human extinction within the next decade. The economic model — by design — does not model this outcome. The researchers' focus is on scenarios where the economy continues to function, which is the appropriate scope for an economic planning tool.

What this means in practice: the model's scenarios should be treated as upper and lower bounds on specific economic variables under specific assumptions. They are tools for stress-testing plans, not oracles.

The AI labor share decline visible even in the modest scenario — from 60% toward 56% — suggests that even slow AI adoption creates distributional pressure on workers relative to capital owners. That finding holds across all scenarios that involve any meaningful AI diffusion. It is arguably the most robust and policy-relevant finding in the entire framework.

Sources: Anthropic working paper PDF | Axios coverage

What Businesses and Workers Should Actually Do With This Information

The Anthropic scenarios converge on a clear strategic implication regardless of which path the economy takes.

Faster AI adoption raises GDP and capital returns while weakening the income share and bargaining power of knowledge workers. This split — growth at the macro level, pressure at the individual worker level — is the defining economic tension of the next four years.

For workers, the key insight is that the jobs most protected in all scenarios are those involving AI-augmented work rather than AI-replaced work. The task-level framework the model uses suggests that professionals who actively integrate AI into their workflows — using it to do more, faster — are in a substantially better position than those who continue performing tasks that AI can now handle independently.

For business owners and operators, the implications are practical and immediate:

  • Organizations that automate repetitive cognitive tasks now reduce their long-term labor cost exposure
  • Companies that build AI-augmented workflows today develop institutional knowledge in using these systems that competitors will struggle to replicate quickly
  • Businesses that rely heavily on manual, task-based knowledge work face growing cost-competitiveness pressure as AI lowers the price of performing those tasks externally

For policymakers, the model's authors and reviewers align around several priorities: expanding access to retraining programs, developing income insurance mechanisms for displaced workers, redesigning tax systems that rely heavily on labor income, and exploring broader capital ownership models that allow workers to share in the returns from AI-driven productivity.

The morning consult survey result — that ordinary Americans broadly expect the substantial scenario of 8.6% GDP growth and moderate unemployment — suggests a public that is cautiously optimistic but not oblivious. Whether that expectation proves accurate depends entirely on choices being made right now in boardrooms, legislatures, and research labs.

Sources: Anthropic Institute econ-scenarios | NPR coverage

Frequently Asked Questions

1. What exactly is the Anthropic AI economy model?

It is the Economic Scenarios for Transformative AI framework, published in September 2026 as Anthropic Institute Working Paper No. 2026-02. It was authored by economists Anton Korinek, Charles I. Jones, Szymon Sacher, Tess Cotter, and Peter McCrory. The model uses the U.S. Department of Labor's O*NET task taxonomy to simulate how different levels of AI capability, adoption, and automation could affect GDP, wages, unemployment, and income distribution in the U.S. economy through 2030. It is a planning tool, not a prediction.

2. What are the three main scenarios and their outcomes?

The modest scenario projects GDP 1.6% above a no-AI path and unemployment rising slightly from 3.8% to 3.9%. The substantial scenario projects GDP 8.3% higher, unemployment at 4.6%, cognitive wages down 0.3%, and labor's income share falling from 60% to 56.1%. The extreme scenario projects GDP 32.4% higher, 15.4% annual growth, 11.9% overall unemployment, 17.9% cognitive worker unemployment, cognitive wages 11.5% below the no-AI path, and labor's income share collapsing to 45.2%.

3. Which scenario do most Americans expect to happen?

Anthropic surveyed 10,980 U.S. adults through Morning Consult from August 11–23, 2026. The median respondent's answers implied an outcome close to the substantial scenario: GDP roughly 8.6% above the no-AI path and overall unemployment around 4.6%.

4. Why are knowledge workers specifically at risk in the higher scenarios?

The model's O*NET task taxonomy identifies management, professional, sales, and office occupations as the most exposed to AI automation. Unlike previous automation waves that displaced physical and routine clerical labor, generative AI can now perform complex cognitive tasks including analysis, writing, legal reasoning, and code generation. In the extreme scenario, cognitive worker unemployment reaches 17.9% — nearly one in five.

5. Does the model predict that AI will cause human extinction?

No. The model explicitly does not include catastrophic AI risk scenarios. It only models scenarios in which the economy continues to function. Anthropic's alignment science lead Evan Hubinger has independently estimated more than a 10% chance of AI causing human extinction within a decade, but that estimate is separate from this economic planning framework and falls entirely outside its scope.

6. What important economic effects does the model leave out?

The model omits policy responses, business cycles, aggregate-demand effects from job losses, financial market disruption, data center buildout impacts, rapid robotics progress, and catastrophic AI risks. It cannot model what happens if AI investment collapses in a bubble, or if mass layoffs trigger a recession through reduced consumer spending.

7. What happens to the labor share of income across all scenarios?

In the modest scenario, labor's share holds close to its 60% baseline. In the substantial scenario, it falls to 56.1%. In the extreme scenario, it falls to 45.2% while capital income rises 81.4% above the no-AI path. This distributional shift — from labor to capital — appears in every scenario involving meaningful AI diffusion, making it one of the model's most robust and policy-relevant findings.

8. When does the model's divergence between scenarios actually begin?

All three scenarios start from the same mid-2026 AI-use readings, where AI currently affects 14% of economy-wide tasks and is used in about 10% of instances within affected tasks. The scenarios begin to meaningfully diverge after 2027, as the differences in assumed capability, adoption speed, productivity, and automation rates compound over time.

9. What do economists say about how accurate these projections are likely to be?

Anthropic economist Anton Korinek notes that while AI capability may advance quickly, economic diffusion could be slower than the models assume. Reviewers including Daron Acemoglu and David Autor flagged concerns about coarse worker heterogeneity, uncertain occupational shrinkage, potentially understated innovation feedback, and missing aggregate-demand effects. The consensus view is to use the scenarios for strategic planning rather than as firm predictions.

10. What practical steps should workers and businesses take based on these findings?

Workers in cognitive roles should actively integrate AI into their current workflows — using it to augment output rather than be replaced by it. Business operators should audit which repetitive cognitive tasks in their operations are automatable now, reducing long-term cost exposure and building competitive institutional knowledge. Policymakers should develop retraining programs, income insurance mechanisms, and capital ownership structures that allow broader sharing of AI-driven productivity gains before disruption accelerates rather than after.

Sources: Anthropic Institute econ-scenarios | Anthropic working paper PDF | NPR coverage | Axios coverage

Additional Links

Flowgrammer.ca – 💡 Unlock AI-Powered Automation! Let flowgrammer.ca handle complex workflows so you can focus on growing your business.

Flowgrammers.ca – 💡 Join Toronto's Fastest Growing AI and Business Automation Community.

OpenClaw Toronto

Episode transcript

Read the full transcript

Craig: imagine an economy getting much richer while your work gets less valuable. anthropic's new economic scenarios put that possibility on the table. the uncomfortable question is who gets paid when machines do more of the work. anthropic, who makes claw, the ai assistants, it's the economist have built a model exploring how ai could affect the american economy through to 2030, and they've published three scenarios from modest change to extreme disruption. these aren't predictions, and the researchers have assigned them probabilities. the biggest question is whether productive gains become better livelihoods. what do you think about, what they've published here, john?

Jon: so i think this is really interesting, and let's just make a distinction here. the headline's kind of loose. right? so anthropics model is primarily a scenario analysis framework, not really a forecast. there's three main cases, modest, substantial, and extreme, and they're not assigned probabilities. so they're just... they are designed to show what follows economically if particular assumptions about ai capability and adoption turn out to be true. now the clever part of the anthropics approach is that it does not simply say ai replaces accountants or ai replaces programmers. instead, what it's doing is it treats occupations as bundles of tasks. so let's let's let's talk about a scenario. consider a nurse. right? a nurse performs dozens of different activities. she or she or they or he documents vital signs, they order supplies, they communicate with patients, they interpret information, coordinate care, all that. so ai may automate some of those activities. they could augment others and leave still others essentially unchanged. so technology may also create completely new tasks such as reviewing ai generated treatment recommendations. right? so antropic breaks task changes into concepts such as automation, where ai performs a task instead of the worker, augmentation, where ai increases the productivity of a human performing the task, new tasks, which technology creates for human workers, and unaffected tasks where ai provides little advantage. so the framework draws on a major strand of modern labor economics associated with economists. so technology doesn't necessarily replace entire occupations in one shot. it just changes which tasks you would perform. and you and i have talked about this many, many times about, related to all these different articles. and so this just kind of takes what we've been talking about and kind of tries to predict what that future might look like. so i guess my question to you is, what do you think about this prediction? and, do you think they did... is this accurate, or what are your thoughts?

Craig: well, i feel like they're hedging their bets. right? because you say prediction. there's three scenarios. right? the modest ai growth, the substantial growth, and the extreme growth. and each of them have different impacts on our society. i guess the question is not what do i think about the prediction, but where in this range do i feel we're gonna land by 2030? now 2030 used to be this, oh, 2030, used to be so far away, but here we are at the end of 2026. yeah. it's three years and a couple of months away. so let's think about that. now do you believe there's gonna be modest growth in this scenario and ai affects 20% of cognitive tasks, 20% diffusion? do you think it's gonna be 30% cognitive tasks with a 40% diffusion and 75% of ai performing task incidents instances, or do you think it's gonna be higher?

Jon: i think it's gonna

Craig: affects 50%. pardon?

Jon: i think it's gonna be higher.

Craig: so you're thinking we're at the extreme ai growth scenario?

Jon: i think so. i... if we continue on the path that we're on right now, i think so. if we slow it down like some... like what some are trying to do, maybe things change. but, as of right now, i think we're on the extreme side. what do you think?

Craig: well, i just don't know if we're smart enough to actually make this change in our businesses and in our economy that fast.

Jon: okay.

Craig: i 2030? well, yeah, it... there's very... it's a very small percentage of companies that are using ai be... beyond, you know, just chatting with it. there's a super small percentage that have implemented automation and some of these tools, but they're getting easier. crockbot, dots looks pretty cool. there's a bunch of other things that are getting easier and easier. so definitely more will happen. mhmm. but we're also gonna get more horror stories. so there'll be... as people adopt these things, they won't put the rules in place. so it'll delete all their emails or it'll do some funky stuff, and then there'll be horror stories so people will back off of it. i think we're gonna be slightly lower than your, the extreme ai growth here only because i think the ai will be capable of it, but i think the humans and the companies will be slow. look at the anti ai settlement right now, the anti data center settlement right now. i think, we're gonna struggle to get the capacity we need, to reach this level, like, the compute capacity. yeah. so i think it's possible. i just think humans gotta change their businesses, change how they run things, change how they do things, then... and, we have to build enough capacity in order to meet that. so i think those are two different things that i don't think will have enough capacity, and i think humans' ability to change won't change fast enough. so i think we're gonna be somewhere stuck in the substantial growth scenario too, which is gonna be interesting. but i guess the question does happen is, you know, this will gonna create a bigger divide between the haves and the have nots, the people who can afford to pay the robots to do the work, to produce stuff, and the people who can't, whether it's robots or ai or what have you. as we've just seen, the divide is getting bigger. like, openai announced a new $500 a month model. so we start to get... we're gonna start to see that up and up and up, and people aren't gonna be able to afford it. so we're gonna see an extreme divide between the rich and the poor. and i'm not sure that we're going to... like, the unemployment and how we pay for all of that, how that's gonna work because our current taxation structure is not ready for this, so something's gotta change.

Jon: yeah. and here's here's what i think though too. and this is... it's kinda funny though. i think white collar professionals could be affected before blue collar workers. and it's... we're already seeing it. like, we're already seeing, like, financial tools and trading tools and all that sort of stuff, are here to be used to to be built. but, a plate fitting ai, is nowhere to be found.

Craig: yeah. so i think you're right where we are all moving to knowledge jobs, as you are white collar jobs. those are all gonna be stripped away because those are the things that ai can do fairly well. when i say all of them, you know, we're still gonna need humans in the loop to some degree, but we're gonna need a lot less of them. what that means exactly, you know, we're gonna place 100 people with, with ai in one person or, not. and we're gonna be moving into a different direction where we're gonna give ai a lot of autonomy, but then there's gonna be regulations and rules around, who could actually make a legal binding and agreement, who is responsible for the deals your a... your ai made, what about lies it told, that kind of... there's a bunch of stuff that we really haven't figured out yet that's gonna slow this down a little bit, but needs to be figured out. and we're really heading

Jon: haven't taken into account yet

Craig: yeah. go ahead. is... what

Jon: if the ai bubble bursts?

Craig: well, and it will. but what does that mean?

Jon: any sort

Craig: of. yeah. you know? you know, the eye bubble will burst because it has to. we... it can't... i... far as i'm concerned, the trillions of dollars they've put into it, there's no way all of that money's coming back. so some of these companies are gonna be winners, and some of them are gonna be losers. there's... they... when the ai bubble bursts as it were, it just means that a few of these company will be less demand. we're getting better open source, better local models, small models, small language models on top of just smaller large language models. and i think the game is changing to what we discussed before where a lot of the compute will start happening again locally, cheap or free. and then the heavy computing, the up to date knowledge, the other stuff you'll send out, and there's gonna be, you know, a good combination of those things. so things are gonna have to change as we go. and the question is, what do we go out for? do we go out for specific lawyer bot? do we go out for medical advice? do we go out... you know, what are those, yeah. high leverage engines that we're gonna go out and pay for? those will be interesting things to see. this is a devastating look at where their job market's gonna go. i'm really concerned about where we're going because, like i said, when the robots do all the work, who gets paid?

Jon: mhmm.

Craig: and it's great that our gdp goes up, but who's buying it? if the people don't have jobs and they can't afford to buy product, it doesn't matter how

Jon: long it's

Craig: going to work.

Jon: would you support a ubi then, craig, in the extreme scenario or no

Craig: well, i'm not saying i don't support ubi. i'm saying it doesn't work. i'm saying someone has to pay for it. so that means our taxation structure has to change, like i was saying earlier, so that we can afford a universal basic income that the government or someone can dole out. the money's gotta come from somewhere. the problem i have is in a society where we remove motivation, then we are in a downward spiral because we try... we attach a lot of our self worth and our... you know, i don't wanna call, i guess, happiness to the value we put in the world, to the work we do when we help somebody, when we build something and whatnot. and when we take that away, it's devastating mentally to people. so i think that our society will go downhill. i just, again, don't know how ubi is gonna work, who's gonna pay for it. so every robot you get, 10% of that robot's income goes... 20% or 50% goes to the government to pay for someone else's living. that could be the case. mhmm. i've... i'm not happy at the thought of that at the moment, but i'm not living in this dystopian world where everyone's unemployed and we need some to solve this problem. so i am happy about solutions. i just... what is... when robots can do the work, what is the value of humans? and what is it we can do that's valuable? it's physical work. it's just not... are we gonna start working for the ais, digging ditches, manufacturing things, fixing things, or doing the awkward things that they just can't do or robots can't do until they can. i mean, it's a bleak picture, but i'm very interested in seeing where it goes. i know you are more positive about ubi than i am. but i'm just saying i don't way to

Jon: put it. yeah.

Craig: i don't know how it works. like, i'm not gonna say you're a fan or this, that you're more positive, but i just... in my mind, i can't see it working with our current economic structures. so if something has to change, and that's gonna be a rough period of time, whatever that change happens to be. yeah. yes. pardon?

Jon: yeah. for sure it will be. yeah. absolutely.

Craig: if ai makes your workplace more productive, what would convince you that workers are sharing the benefit? the real impact of this work is making the dispute... distribution questions harder to ignore. growth creates possibilities, but sharing the gains require choices. a richer economy should be something people experience, not something, chart announces. so in the... let us know in the comments what specific change would make ai predict productivity worthwhile in your workplace. we'd love to hear from you.

More from Silicon Scoop

All episodes