The Rise of Artificial Intelligence: Understanding Its Impact and Potential

On this page

1. What Is Artificial Intelligence?
2. The Economic Impact of AI
3. How AI Learns: The Foundations
4. Understanding the Spectrum of AI Technologies
5. Artificial Narrow Intelligence (ANI)
6. Generative AI
7. Agentic AI
8. Artificial General Intelligence (AGI)
9. The Pace of Adoption
10. Navigating the Gap Between Hype and Reality
11. Key Components of AI Implementation
12. Data Quality and Governance
13. Measuring Success
14. Societal Implications of AI
15. Economic Divergence Between Nations
16. Bias in AI Systems
17. Workforce Evolution
18. Divergent Economic Projections
19. The Cost and Infrastructure Challenge
20. Building AI Capabilities in Organizations
21. Conclusion
22. Glossary
23. References
24. Footnotes

AI is projected to contribute up to $15.7 trillion to global GDP by 2030, according to combined research from McKinsey, Gartner, PwC, and other leading institutions.¹ Goldman Sachs estimates that generative AI alone has the potential to improve productivity growth by 1.5% and raise global GDP by 7%, the equivalent of $7 trillion, over the next decade.² McKinsey’s baseline sizing for generative AI remains between $2.6 and $4.4 trillion of annual value across 63 identified use cases, with the bulk concentrated in customer operations, marketing and sales, software engineering, and research and development.³

What Is Artificial Intelligence?#

Before examining AI’s impact, it helps to define the term itself. Artificial intelligence is the ability of a computer system to perform tasks that normally require human intelligence: things like recognizing speech, identifying objects in images, making predictions, or generating text. Rather than following a fixed set of instructions written by a programmer for every possible situation, modern AI systems learn patterns from large amounts of data and apply those patterns to new situations.

The reason this matters is profound: traditional software can only do exactly what it was explicitly told to do, line by line. AI, by contrast, can handle situations its creators never anticipated, which is why it can be applied to messy, real-world problems, like understanding a customer’s oddly worded question, that were previously impossible to automate.

The Economic Impact of AI#

The global AI market has reached remarkable scale, valued at $391 billion as of 2025 and projected to reach $1.81 trillion by 2030.⁴ This expansion is being driven by increasing enterprise adoption, rising consumer AI interactions, and public-sector integration, growing at a pace faster than the cloud computing boom or the mobile app economy of the 2010s. Enterprise generative AI spending alone reached $13.8 billion in 2024, six times the $2.3 billion spent in 2023,⁵ while Goldman Sachs estimates global AI investment will reach $200 billion by 2025.⁶

The software industry has already realized significant benefits from AI implementation, but future growth extends far beyond the tech sector. In retail, for example, the AI market was valued at $7.14 billion in 2023 and is projected to grow at a 31.8% compound annual growth rate (CAGR) to reach $85.07 billion by 2032.⁷ Industries such as travel, transportation, automotive, materials, and manufacturing are similarly positioned to experience substantial AI-driven transformations. AI’s reach is so broad because nearly every industry runs on data and decision-making, and AI excels at both. So wherever there is information to analyze or a repetitive judgment to make, AI can add value.

How AI Learns: The Foundations#

Most of today’s AI is powered by machine learning, a method in which a computer learns to perform a task by studying examples rather than being explicitly programmed. A useful analogy: instead of writing thousands of rules to describe what a cat looks like, you show the system thousands of labeled photos of cats and non-cats, and it gradually figures out the distinguishing patterns on its own. This is why data quality is so critical. If the examples are flawed or biased, the system learns the wrong lessons.

The most common form is supervised learning, where the AI is trained on data that has been labeled with the correct answers, for example emails marked “spam” or “not spam.” The system learns to map inputs to outputs, so it can then classify new emails it has never seen. This technique powers much of AI’s real-world value today, from fraud detection to medical image analysis.

A major driver of recent breakthroughs is deep learning, a subset of machine learning that uses neural networks, computing systems loosely inspired by the way neurons connect in the human brain. These networks are built from layers of interconnected nodes that pass information forward, with each layer detecting increasingly complex features. The reason deep learning has been so transformative is that, given enough data and computing power, it can automatically discover subtle patterns that humans could never hand-code, which is exactly what makes tasks like language translation and voice recognition possible.

Understanding the Spectrum of AI Technologies#

Artificial intelligence encompasses distinct categories that serve different purposes:

Artificial Narrow Intelligence (ANI)#

The majority of practical AI applications today fall under artificial narrow intelligence (ANI), systems designed to perform one specific task, such as voice recognition in smart speakers, autonomous vehicle navigation, or agricultural monitoring. They are sometimes called “one-trick” systems because they cannot transfer their skill to unrelated tasks; a self-driving car’s AI cannot suddenly write an email. Yet when the single trick is a valuable one, these systems can be extraordinarily useful, which is why ANI accounts for the vast majority of AI’s economic impact today.

Generative AI#

The recent emergence of generative AI represents a significant advancement toward more versatile systems. Where earlier AI mostly classified or predicted, generative AI creates: new text, images, audio, or code that did not exist before. This is why it feels more general-purpose: a single generative model can act as a copy editor, brainstorming partner, and text summarizer. AI copilots have been shown to improve knowledge-worker productivity by 30 to 45%.⁸ McKinsey reported that regular generative AI use reached approximately 65% of organizations in early 2024, a rapid jump from 2023’s early adoption phase.⁹

Agentic AI#

The next productivity frontier is agentic AI, systems that can plan, execute, and adapt autonomously to complete complex, multi-step tasks with minimal human intervention. The distinction matters because most AI to date has assisted a human doing a task, whereas an agent can carry out an entire workflow on its own, such as researching a topic, drafting a report, and scheduling follow-up actions. In McKinsey’s 2025 survey, 23% of respondents reported their organizations are scaling an agentic AI system, with an additional 39% experimenting with them.¹⁰ If you want to try one rather than read about them, Taskade (affiliate link) is one of the platforms that lets non-programmers build and run agentic workflows of this kind.

Artificial General Intelligence (AGI)#

Artificial general intelligence (AGI) is the theoretical concept of AI capable of performing any intellectual task a human can, potentially even exceeding human ability across all domains. The key difference from everything above is generality: an AGI would not be limited to one trick or one type of content but could reason flexibly across completely unrelated problems. While progress in ANI and generative AI has been substantial, AGI remains a distant goal that will likely require multiple technological breakthroughs and may be decades or even centuries away.

Check yourself

Three quick scenarios. Match each to the right point on the spectrum.

The Pace of Adoption#

AI has achieved near-universal business adoption, with 88% of businesses reporting they use AI in at least one capacity in 2025, a dramatic acceleration from 78% in 2024 and 55% in 2023.¹⁰ Employee-level adoption is climbing in step: Gallup found that 27% of white-collar workers now frequently use AI at work, an increase of 12 percentage points from 2024.¹¹

Investment continues to climb sharply. Projected private U.S. investment in AI is expected to grow to $81.7 billion in 2025, up from $47.4 billion in 2022.¹² Nearly all companies plan to invest further, with 92% intending to increase AI investments over the next three years according to McKinsey’s 2025 State of AI report.¹³

Navigating the Gap Between Hype and Reality#

The rapid advancement in narrow AI and generative systems has created both excitement and unnecessary hype. While the progress is real, a realistic understanding of AI’s current limitations remains essential. Despite widespread individual adoption, structural economic effects have so far appeared somewhat limited; the number of U.S. businesses formally using AI rose from 3.7% in September 2023 to only 5.4% by February 2024.¹⁴ The gap exists because experimenting with a tool is easy, but rebuilding entire business processes around it is hard.

Firms reporting meaningful revenue impacts from AI are considerably more likely to have fully scaled their use rather than confining it to basic applications.¹⁰

The 2025 State of AI research reflects a maturing market: more teams use AI daily, but most firms still struggle to scale it and prove enterprise-level impact.¹⁰ Leaders stand out by setting growth objectives, rewiring workflows, and establishing measurable controls, often with AI agents front and center.¹⁰

Key Components of AI Implementation#

Data Quality and Governance#

The effectiveness of AI depends heavily on the quality of the data used to train it. This is because AI learns patterns directly from its data. If that data is incomplete, outdated, or biased, the AI faithfully reproduces those flaws, a principle often summarized as “garbage in, garbage out.” Organizations are increasingly advised to govern for both safety and speed by establishing an approvals matrix by risk tier, pre-approving trusted tools and datasets, and logging prompts and outputs to shorten time-to-value while satisfying compliance requirements.¹⁵

Measuring Success#

Successful AI implementation requires tracking both leading and lagging indicators. These include adoption metrics such as active users and tasks automated; quality metrics such as hallucination rates and guardrail blocks; and business results such as EBIT impact and revenue lift.¹⁵ Tracking these matters because widespread usage alone does not guarantee value; only measurable outcomes reveal whether AI is actually helping the business.

Societal Implications of AI#

Economic Divergence Between Nations#

AI is reshaping global economies, affecting developed and developing nations differently. Because occupational structures differ across countries, the share of “high-exposure” jobs, roles whose tasks AI can readily perform, varies substantially, reaching up to 60% in advanced economies versus 40% in emerging markets and just 26% in low-income countries.¹⁶ This divergence happens because wealthier economies have more knowledge-and-office work, which is precisely what today’s AI is best at automating.

Adoption is nonetheless accelerating in developing economies. Still, a McKinsey survey spanning 105 countries found that while most companies now use AI, an overwhelming majority have not moved beyond basic applications, especially among small and medium-sized businesses.¹⁰

Bias in AI Systems#

Because AI learns from historical data, it can absorb and even amplify the biases embedded in that data. For example, an AI trained on past hiring decisions may learn to favor the same groups those decisions favored, perpetuating discrimination at scale. This is why identifying, measuring, and mitigating bias has become a central concern of responsible AI deployment. An unchecked model can quietly automate unfairness while appearing perfectly objective.

Workforce Evolution#

The impact of AI on employment varies across industries and job categories. The World Economic Forum’s Future of Jobs Report 2025 revealed that 41% of companies worldwide expect to reduce their workforce by 2030 due to AI automation.¹⁷ At the same time, AI is projected to create 97 million new jobs globally as entirely new categories of work emerge.¹⁸ Goldman Sachs estimates that generative AI could expose the equivalent of 300 million full-time jobs worldwide to automation.²

Importantly, AI’s current role in the workplace remains primarily to assist workers and streamline workflows rather than to fully replace human jobs. Human oversight remains essential because generative AI can still produce errors or hallucinations, so accuracy-critical work requires a person to review its output. As a proof point, developers at Goldman Sachs have written as much as 40% of their code automatically using generative AI during proof-of-concept stages.¹⁹

Divergent Economic Projections#

Economists remain divided on the ultimate scale of AI’s impact. The Council of Economic Advisers has noted the wide range of long-run GDP estimates, spanning McKinsey’s projection of a 2.4 to 4.1% increase to Goldman Sachs’ estimate of a 7% increase after ten years.²⁰ More conservative analyses caution that anticipated productivity gains may be modest once challenging, context-dependent tasks are considered, while highlighting potential inequality dynamics tied to AI-driven automation.²¹

The Cost and Infrastructure Challenge#

The computational demands of cutting-edge AI have grown enormously. From 2016 to 2024, the energy and amortized hardware costs to train an AI model grew at an average rate of 2.4 times per year, while cloud compute costs grew at an average rate of 2.5 times per year.²⁰ This matters because nearly a decade of annually doubling costs concentrates frontier AI development in the hands of a few well-resourced organizations and nations, those who can afford the enormous computing bills required to train the largest models.

Building AI Capabilities in Organizations#

Companies seeking to leverage AI effectively should consider the following approaches:

• Develop AI transformation strategies tailored to organizational needs
• Build specialized AI teams with cross-functional expertise
• Establish data governance and risk-tier approval processes
• Track both adoption and business-outcome metrics
• Foster AI literacy across all organizational levels

Conclusion#

Artificial intelligence represents a transformative technology with far-reaching implications for business and society. In 2025, AI is no longer a concept in the future tense. It is a defining force of the present, evolving faster than any prior technological revolution and becoming embedded into everything from consumer services and industrial operations to national defense and scientific discovery. By understanding its realistic capabilities and limitations, what it is, how it learns, and why it succeeds or fails, organizations can effectively harness AI’s potential while avoiding common pitfalls. The central challenge of the coming years is no longer whether to adopt AI, but how to scale it responsibly and translate widespread usage into measurable, enterprise-level value.

Reference

Glossary#

CAGR

Compound annual growth rate: the steady year-over-year percentage at which a value grows over a period of time.

EBIT

Earnings before interest and taxes, a standard measure of a company's operating profit.

Hallucination

When a generative AI produces a confident but factually incorrect answer.

Guardrail

A safety filter that blocks an AI from producing unwanted, unsafe, or off-policy output.

ANI

Artificial narrow intelligence: AI built to do one specific task well, unable to transfer its skill to unrelated tasks.

AGI

Artificial general intelligence: the still-theoretical concept of AI that could perform any intellectual task a human can, across unrelated domains.

High-exposure job

A role whose day-to-day tasks today's AI can readily perform, and which is therefore most affected by automation.

Sources

References#

Market-size, adoption, and forecast figures below are 2024–2026 snapshots from the cited primary sources.

Footnotes#

1. PwC, “Sizing the Prize” (Global Artificial Intelligence Study), 2017. pwc.com/sizing-the-prize. ↩

2. Goldman Sachs, “Generative AI could raise global GDP by 7%,” 5 Apr 2023. goldmansachs.com/insights. ↩ ↩²

3. McKinsey, “The economic potential of generative AI: the next productivity frontier,” Jun 2023. mckinsey.com/insights. ↩

4. Grand View Research, “Artificial Intelligence Market Size, Share & Trends” (2025–2030 edition), 2025. grandviewresearch.com. ↩ ↩²

5. Menlo Ventures, “2024: The State of Generative AI in the Enterprise,” 20 Nov 2024. menlovc.com. ↩

6. Goldman Sachs Research, “AI investment forecast to approach $200 billion globally by 2025,” 1 Aug 2023. goldmansachs.com/insights. ↩

7. Fortune Business Insights, “Artificial Intelligence (AI) in Retail Market” (2024 edition), 2024. fortunebusinessinsights.com. ↩

8. Dell’Acqua et al. (Harvard / BCG field experiment), “Navigating the Jagged Technological Frontier,” Sep 2023. bcg.com/publications. ↩

9. McKinsey, “The state of AI in early 2024: gen AI adoption spikes and starts to generate value,” May 2024. mckinsey.com/insights. ↩

10. McKinsey, “The state of AI in 2025: agents, innovation, and transformation,” Nov 2025. mckinsey.com/insights. ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷

11. Gallup, “AI Use at Work Has Nearly Doubled in Two Years,” Jun 2025. gallup.com/workplace. ↩

12. Goldman Sachs projection, via Congressional Research Service, “The Macroeconomic Effects of Artificial Intelligence” (IF12762), 2024. congress.gov. ↩

13. McKinsey, “Superagency in the workplace: empowering people to unlock AI’s full potential,” Jan 2025. mckinsey.com/insights. ↩

14. U.S. Census Bureau, Bonney et al., “Tracking Firm Use of AI in Real Time” (CES-WP-24-16), 2024. census.gov. ↩ ↩²

15. McKinsey, “Implementing generative AI with speed and safety,” 2024. mckinsey.com/insights. ↩ ↩²

16. IMF, Cazzaniga et al., “Gen-AI: Artificial Intelligence and the Future of Work” (Staff Discussion Note SDN/2024/001), 14 Jan 2024. imf.org/blogs. ↩ ↩²

17. World Economic Forum, “The Future of Jobs Report 2025,” Jan 2025. reports.weforum.org. ↩

18. World Economic Forum, “The Future of Jobs Report 2020,” Oct 2020. weforum.org. ↩

19. Marco Argenti (Goldman Sachs CIO), interview reported by CNBC, 22 Mar 2023. cnbc.com. ↩

20. U.S. Council of Economic Advisers, “Artificial Intelligence and the Great Divergence,” Jan 2026. whitehouse.gov. ↩ ↩² ↩³

21. Daron Acemoglu, “The Simple Macroeconomics of AI,” NBER Working Paper 32487, May 2024. economics.mit.edu. ↩