AGI and the Future of Growth, Work, and Power

Artificial general intelligence (AGI)  AI systems that can perform a very wide range of intellectual tasks at or above top human expert level  would not merely add another technology to todays economy. It could fundamentally reorganize how production, wealth, and power operate. This article, inspired by an in-depth interview with researcher Carl Shulman, examines what might happen to growth, work, inequality, and national security if general-purpose AI becomes as capable and common as electricity.

From Gradual Improvement to Economic Phase Change

Modern growth has been driven by incremental innovation: better machines, improved management, and steady scientific discovery. AGI could compress centuries of such progress into a few years by transforming two central inputs to growth: cognitive labor and physical labor.

• Cognitive work (research, management, design, legal analysis, software engineering) could be automated and scaled almost without limit once it runs on hardware instead of human brains.
• Physical work could be increasingly performed by robots guided by advanced AI, removing many of todays bottlenecks around training, safety, and coordination.

Traditionally, world GDP per capita has grown at a few percent per year, roughly doubling every few decades. With AGI, some economists and AI researchers argue that growth rates could temporarily rise by orders of magnitude if capital, labor, and knowledge all become far more scalable.

Why Economic Growth Could Explode After AGI

Several reinforcing mechanisms could drive a dramatic acceleration in growth once AGI is available.

1. Turning Compute into Talent

Today, highly skilled workers such as leading engineers or scientists are scarce and expensive. In an AGI world:

• Training an additional AI worker mainly requires hardware and energy, not 20 years of human education.
• Copies can be made quickly, allowing one core design to be instantiated millions of times.
• Once trained, AGI workers do not retire and can operate continuously.

If the marginal cost of world-class professional labor falls by orders of magnitude, firms and governments would have strong incentives to expand AI-based workforces as quickly as hardware can be manufactured.

2. Feedback Loops in AI Progress

AGI systems could themselves be applied to AI research.

• AI engineers, system designers, and theoreticians could be replaced or assisted by AGIs, increasing the rate at which better models and training methods are discovered.
• This can create a self-reinforcing loop: better AI leads to faster AI research, which leads to even better AI.
• Such dynamics are already visible in narrower settings, where machine learning helps design chips and optimize software pipelines.

While practical constraints like data availability and hardware limits would eventually slow this loop, there may be a period where each improvement rapidly enables the next.

3. Removing the Human Bottleneck in Physical Production

Even in capital-rich countries, many industries are constrained by human availability, training time, or safety issues. Robots guided by advanced AI could perform tasks that currently require dexterous, adaptive human workers, including:

• Construction, logistics, and advanced manufacturing
• Agriculture and food processing
• Maintenance and infrastructure operations

Industrial robotics is already widely used in automotive and electronics manufacturing, increasing productivity and output. If general robotic manipulation and on-site problem-solving become reliable and cheap, production lines, warehouses, and construction sites could be scaled with far fewer human workers, constrained mainly by materials and energy.

Why We May Not See Clear Warning Signs in Today’s Data

One skeptical argument is that if AGI were close, we would already see accelerating growth from today’s AI and digital technologies. Instead, many advanced economies have experienced relatively modest productivity growth in recent decades.

There are several reasons this pattern may be compatible with a future discontinuous jump:

• Threshold effects: Many tasks remain out of reach for current systems, particularly those needing robust general reasoning, long-term planning, or flexible physical interaction. Large benefits may only appear once systems cross a certain capability threshold.
• Slow diffusion of IT: It historically took years or decades for earlier general-purpose technologies – like electricity or computers – to fully transform production processes and show up in macro statistics.
• Under-measured digital value: Some economic value from software and online services is not fully captured in GDP, especially when services are provided for free or nearly free.

AGI could trigger a qualitatively different regime from today’s incremental digitalization if it can perform the full range of tasks needed to design, implement, and manage new processes at scale.

Work, Wages, and Distribution in an AGI World

AGI is likely to reshape labor markets more profoundly than past waves of automation. The key issues are not only job loss but also who ultimately controls and benefits from extremely capable AI.

Will Human Workers Be Completely Displaced?

Economic theory indicates that when machines can substitute for labor across a wide range of tasks, the demand for human labor can fall dramatically unless new complementary activities emerge.

• In scenarios where AGI can do nearly all tasks cheaper and better, most current occupations could vanish over time.
• Some roles might remain for preference or legal reasons (e.g., human-facing services, symbolic leadership, politics), but these may represent a small share of total hours worked.
• New creative or interpersonal jobs could arise, but AGI systems might also be competitive in many of these areas.

However, mass unemployment does not automatically imply mass poverty. With enough overall wealth and political will, societies could decouple income from traditional employment.

How Might People Be Supported Without Jobs?

High-income countries already use taxes and transfers to redistribute a significant share of economic output to those not directly producing goods and services – including retirees, children, and recipients of social programs.

If AGI causes total output to increase by a factor of 100 or 1,000, even unchanged redistribution mechanisms could support far higher material living standards for non-working citizens. Potential approaches include:

• Expanded social insurance funded by taxes on AI-intensive industries, capital income, or resource rents.
• Universal or near-universal transfers that provide a basic standard of living independent of employment.
• Public ownership stakes in key AI infrastructure or natural resources, turning large-scale automation into a shared dividend.

The feasibility of such arrangements depends heavily on institutions, politics, and global coordination, not just on technical capacities.

Inequality and Ownership of AGI Systems

Who owns and controls AGI could matter as much as the technology’s raw capabilities. Several dimensions of inequality are especially relevant.

Concentration of Capital and Intellectual Property

Large-scale AI development today is concentrated in a small number of firms and governments that can afford the necessary data centers, research teams, and compute budgets. Existing trends toward market concentration in digital industries suggest that, absent deliberate policy, AGI profits could be highly concentrated.

Mechanisms that could amplify this include:

• Economies of scale in training and deployment of advanced models.
• Network effects where dominant platforms attract more users, data, and complementary services.
• Patents, trade secrets, and proprietary data that limit competition.

Highly concentrated control of AGI could translate into extreme wealth inequality and significant political influence for a small number of actors.

Geopolitical Inequality Between Nations

At the international level, countries that first achieve and secure advanced AGI capabilities could enjoy large strategic advantages, including:

• Far higher economic output per capita.
• Leadership in military and intelligence applications of AI.
• Control over critical hardware supply chains, such as advanced semiconductor manufacturing.

Existing imbalances in research capacity, compute infrastructure, and chip manufacturing raise the risk that AGI amplifies current power disparities rather than closing them.

National Security in a World of AGI

AGI has obvious implications for national security, stretching from cyber operations to strategic stability between great powers.

AI-Driven Military and Intelligence Capabilities

Militaries and security agencies are already investing heavily in AI for intelligence analysis, logistics, and autonomous systems. AGI could multiply these capabilities:

• Planning and analysis: Faster and more thorough scenario modeling, war-gaming, and operational planning.
• Cyber operations: Highly capable agents that can discover software vulnerabilities, design sophisticated malware, and defend critical infrastructure.
• Autonomous platforms: Swarms of air, sea, and land systems that can coordinate with minimal human supervision.

This could alter traditional balances of power by favoring actors who can best integrate AGI into their command, control, and manufacturing systems.

Risks of Destabilization and Arms Races

Strategic competition over AGI could create several destabilizing dynamics:

• Racing to deploy: States may fear that delaying safety measures could allow rivals to gain a decisive advantage, leading to underinvestment in testing and alignment.
• Opacity: The internal workings of complex AI systems may be difficult to interpret, complicating verification of agreements and raising mistrust.
• Lower barriers to powerful weapons: Automated design and optimization tools could speed progress in areas like cyber weapons and possibly biological threats, if not carefully controlled.

Managing these risks will likely require new international norms and possibly inspection or monitoring regimes for high-end AI development, analogous to existing arms control arrangements.

Constraints: Energy, Hardware, and Physical Limits

Growth cannot accelerate indefinitely. Even in an AGI scenario, real-world constraints will shape what is possible.

Energy and Compute Limitations

Running vast numbers of AGI systems and general-purpose robots would require significant energy and computing infrastructure.

• Data centers are already major electricity consumers, and AI workloads are projected to increase that demand substantially.
• Advanced semiconductor manufacturing is geographically concentrated and technically challenging, which creates bottlenecks and strategic vulnerabilities.
• Thermodynamic and information-theoretic limits ultimately cap how much computation can be extracted from a given energy supply, although current systems are far from those limits.

In the medium term, expanding renewable energy generation, transmission networks, and cooling systems would be critical for scaling AGI use without unacceptable environmental impacts.

Physical and Environmental Constraints

Even with AI-directed automation, the physical world cannot be reshaped arbitrarily fast.

• Mining, materials extraction, and manufacturing still require time, logistics, and environmental safeguards.
• Climate change, biodiversity loss, and pollution place limits on how much and how quickly energy use and industrial output can grow without severe ecological consequences.

AGI could help design more efficient processes and low-carbon technologies, but it could also accelerate harmful exploitation if governance is weak.

Policy Priorities for an AGI-Transformed Future

Governments, firms, and civil society have a narrowing window to shape institutions and norms around advanced AI. Several priorities emerge from the economic and security analysis.

1. Safety and Alignment Research

As AI systems become more capable and autonomous, ensuring they act in accordance with human intentions becomes increasingly important. Official reports have emphasized the need for substantial investment in AI safety research, robust testing, and monitoring of powerful systems.

2. International Coordination and Guardrails

Because AGI development and deployment have cross-border consequences, unilateral approaches are unlikely to be sufficient.

• Major powers may need to agree on basic norms governing deployment of AI in military contexts, including constraints on fully autonomous lethal systems and destabilizing cyber operations.
• Shared standards for auditing high-risk models could reduce the chance of catastrophic misuse.
• Cooperation on export controls and monitoring of advanced chips could help prevent uncontrolled proliferation while preserving peaceful scientific use.

3. Domestic Economic Institutions

To make an AGI-driven economy broadly beneficial, countries may need to rethink taxation, social insurance, and competition policy.

• Tax policy might shift toward capital, resource rents, or AI outputs rather than labor income.
• Antitrust enforcement could be critical to prevent excessive concentration of AGI capabilities in a small number of firms.
• Education and retraining will still matter, particularly during transition periods before full automation is possible.

Frequently Asked Questions (FAQs)

Q1: Does AGI guarantee that everyone becomes much richer?

No. While AGI could greatly expand total output, how that wealth is distributed depends on institutions, ownership, and policy. Without careful design, gains could be heavily concentrated among a small number of firms or states.

Q2: Will there still be any meaningful work for humans?

It is likely that many tasks humans do today will be automated, especially routine or predictable work. Some roles may remain due to social preference, legal requirements, or uniquely human traits, but total demand for human labor could fall sharply if AGI is broadly capable.

Q3: How soon could these transformations happen?

Expert forecasts about when AGI might be developed vary widely, and there is no consensus. Some researchers expect substantial advances within a few decades, while others are more skeptical. The economic and security implications described here are conditional on AGI actually being achieved.

Q4: Is there anything individuals can do to prepare?

Individuals can focus on skills that are complementary to AI – such as governance, ethics, systems thinking, and institution-building – and support evidence-based policy efforts around AI safety, transparency, and benefit-sharing.

Q5: Could AGI help solve global challenges like climate change?

Potentially yes. AGI could accelerate the discovery of new materials, energy systems, and mitigation strategies, as well as optimize existing infrastructure. But if deployed irresponsibly, it could also speed up resource use and emissions. Outcomes will depend on how the technology is steered.

Written by

Sneha Tete

• Integrated M.A. in Applied Linguistics, University of Hyderabad
• 5+ years in editorial, instructional, SEO, and research-based writing

Sneha is a relationships and lifestyle writer with a strong foundation in applied linguistics and certified training in relationship coaching. She brings over five years of writing experience to theskillvault,  crafting thoughtful, research-driven content that empowers readers to build healthier relationships, boost emotional well-being, and embrace holistic living. Her writing and insights reflect the belief that beauty, happiness, and wellness…