Anthropic: AI Has Not Led to an Increase in Unemployment Rate

marsbitPublicado a 2026-08-24Actualizado a 2026-08-24

Resumen

Anthropic's Chief Economist, Peter McCrory, argues that despite significant growth in AI adoption, it has not yet led to increased unemployment in the US. Current data shows stable unemployment rates, even in occupations highly exposed to AI automation. His analysis, based on Anthropic's research over 18 months, suggests AI is currently a "skill-biased, labor-augmenting" technology. It enhances human capabilities rather than replacing workers because every job contains tasks—like interpersonal coordination or physical interaction—that AI cannot systematically perform. Furthermore, effective use of AI requires high user expertise, and human roles in planning and judgment remain crucial, even as agents handle more delegated tasks. McCrory acknowledges three future scenarios—software, economic, and Coasean singularities—where AI could recursively improve itself or drastically lower transaction costs, potentially disrupting labor markets and organizational structures. However, he posits that "weak links," the indispensable tasks resistant to automation, will act as a buffer, preventing a collapse in labor's share of income. He predicts US unemployment will not rise significantly due to AI within the next 12 months, with key indicators being unemployment in AI-exposed jobs and the return on human expertise in agent-based workflows.

AI has not taken anyone's job. At least not so far.

This conclusion comes from a lengthy article recently published by Peter McCrory, Head of Economic Research at Anthropic.

https://x.com/PeterMcCrory/article/2079979321607745905

Quality-adjusted AI output grew over 2000% in both 2024 and 2025, and about one-fifth of US businesses already use AI in their operations, yet the US unemployment rate in June remained at 4.2%, right at the full employment level.

McCrory's core judgment is: AI is currently still a "skill-biased, labor-augmenting" technology; it is amplifying human capabilities, not replacing humans.

He also gave this judgment an expiration date: 12 months.

Data Doesn't Lie: Even High AI-Exposure Jobs Are Fine

McCrory compiled a full chain of data to demonstrate how stable the US labor market is: unemployment rate at 4.2%, the job openings-to-unemployed ratio just above 1, employment-population ratio for ages 25-54 near multi-decade highs, and initial jobless claims running low for four consecutive years.

He believes that, overall, the US employment situation is stable and improving.

Aggregate data is just the starting point; cross-group comparisons reveal the real picture.

McCrory and his colleague Maxim Massenkoff previously conducted an analysis: they isolated jobs corresponding to tasks that Claude is heavily used to automate and examined whether their unemployment rates were deteriorating faster than other jobs.

The answer was no.

Updating with the latest Bureau of Labor Statistics data, the conclusion remains unchanged.

Job postings for software engineering positions have rebounded since May 2025, with growth rates even exceeding the average for other positions.

The only group worth noting is young workers.

McCrory acknowledges that it has indeed become harder for young people to find jobs in high AI-exposure positions, and a paper from Stanford's Digital Economy Lab provides similar evidence.

But he also points out an easily overlooked context: since 2022, the US has experienced the largest non-recessionary labor market cooling on record. The "low hiring, low layoffs" pattern naturally squeezes early-career individuals first.

Attributing this entirely to AI lacks sufficient evidence.

Why Jobs Haven't Been Taken: Every Job Has Parts AI Can't Handle

McCrory's explanation is built on the factual foundation of multiple Anthropic research reports from the past 18 months, with a logical chain as follows:

First-layer fact: Among all occupational classifications in the US Department of Labor's O*NET, there is not a single job where all associated tasks are systematically completed by Claude.

Every job has components that AI struggles with: interpersonal coordination, face-to-face interaction, and dealing with the physical world.

These components act like the short planks of a barrel, both limiting the extent to which AI can boost overall productivity and making people who can fill these gaps more valuable.

Second-layer fact: The stronger AI becomes, the higher the requirements for the people who operate it.

Anthropic's Economic Primitives report from January this year found that the more complex Claude's output, the higher the professional expertise of the user's input behind it.

When the model produces a complex economic model, it's often a professional user capable of giving precise instructions who is directing it.

https://www.anthropic.com/research/economic-index-primitives

The Learning Curves report in March showed that after six months of using Claude, users increasingly employed it as a thinking partner, with higher interaction success rates.

If AI were already strong enough to perform jobs independently, this learning effect shouldn't occur.

https://www.anthropic.com/research/economic-index-march-2026-report

Third-layer fact: Even in the era of Agents, human value remains solid.

Anthropic analyzed seven months of Claude Code usage data and found that task value continued to climb, but the return on human professional skills did not decline:

Users are responsible for planning and delegate implementation to Claude;

Users with stronger professional skills have higher task success rates and recover faster when Claude makes mistakes.

The premium for pure programming implementation might be shrinking, but the value of planning, judgment, and error correction is rising instead.

An interesting piece of supporting evidence comes from users' own perceptions.

In Anthropic's June Economic Index survey, over one-third of respondents expected AI to be able to do most of their job within 12 months, but the proportion expecting to become unemployed as a result was significantly lower.

Users with higher levels of AI automation were even more optimistic about their salaries and job security.

https://www.anthropic.com/research/economic-index-june-2026-report

Three Types of "Singularities" and One Line of Defense

McCrory's assessment of the present is optimistic, but he does not avoid discussing three possible future scenarios that could disrupt the current landscape.

They are referred to in academic literature and industry discussions as the Software Singularity, the Economic Singularity, and the Coasean Singularity.

https://x.com/MTSlive/status/2090610614154764509

The Software Singularity refers to the automation of innovation itself by AI.

No matter how powerful past general-purpose technologies were, an internal combustion engine couldn't invent new modes of transportation.

But AI is directly endowed with general cognitive abilities, theoretically capable of accelerating its own iteration in return.

McCrory's article explicitly discusses this scenario: if recursive self-improvement (RSI) becomes a reality, the improvement of AI capabilities will break free from the pace of human control.

The Economic Singularity is the economic consequence of the Software Singularity.

McCrory cites a 2017 model by Aghion, Jones, and Jones, pointing out that if the innovation process itself is automated, standard economic models predict "infinite growth in finite time."

This conclusion sounds like science fiction, but it is the result of rigorous mathematical deduction.

The Coasean Singularity targets another dimension.

An NBER paper from 2025 (co-authored by Shahidi, Rusak, Manning, et al.) proposed that when Agents reduce transaction costs like search, negotiation, contracting, and supervision to near zero, Ronald Coase's classic 1937 question is reactivated—once the cost of market coordination disappears, the very reason for a firm's existence is called into question.

https://www.nber.org/system/files/chapters/c15309/c15309.pdf

The Agent scenario discussed by McCrory naturally dialogues with this framework: if Agents not only change production methods but also organizational forms, the impact on the labor market will be far more complex than simply "who lost their job."

Faced with the three singularities, McCrory offers the same line of defense: the "weak link."

This concept comes from the paper by Aghion, Jones, and Jones—the ceiling of economic growth is not determined by the best-performing links, but by those indispensable links that are difficult to improve.

https://www.nber.org/system/files/working_papers/w23928/w23928.pdf

As long as there exist critical tasks that consistently cannot be automated, whether due to technical bottlenecks or social constraints, the labor income share will not collapse, and corporate forms will not completely dissolve.

The current state of coding Agents is already validating this logic.

Data cited by McCrory shows that after introducing coding Agents, lines of code increased by 10 to 20 times, but the number of software releases only increased by 30%, with no growth in the actual usage of applications.

Coding is just one part of software production; links like testing, architectural design, and product judgment are still in human hands—they are the current "weak links."

McCrory gives a clear prediction: One year from now, the US unemployment rate will not have risen significantly due to AI.

This judgment is valid for 12 months, and the signals for verification are clear—observe whether the unemployment rate for high AI-exposure positions begins to diverge from other positions, and whether the return on human professional skills in Agent scenarios begins to decline.

The day both indicators turn simultaneously is the day the analytical framework he built himself will also need to be revised.

References:

https://x.com/PeterMcCrory/article/2079979321607745905

https://x.com/MTSlive/status/2091252532689658001

This article is from the WeChat public account "新智元" (New Zhiyuan), author: ASI启示录 (ASI Apocalypse), editor: 马可 (Ma Ke)

Preguntas relacionadas

QAccording to the article, what is Anthropic's economic research director Peter McCrory's main conclusion about AI's impact on employment so far?

AHis core conclusion is that AI currently remains a 'skill-biased, labor-augmenting' technology. It is amplifying human capabilities rather than replacing people.

QWhat key data point does McCrory use to argue that the U.S. job market remains stable despite widespread AI adoption?

AHe points out that the U.S. unemployment rate in June was 4.2%, which is exactly at the full employment level, and that even for high-AI-exposure occupations, the unemployment rate has not deteriorated faster than other jobs.

QWhat are the 'weak links' that McCrory identifies as a key reason why AI has not yet caused widespread job loss?

AThe 'weak links' are tasks within every job that AI cannot handle systematically. These include interpersonal coordination, face-to-face interaction, and tasks involving interaction with the physical world. They limit AI's overall productivity gains and make people who can handle these tasks more valuable.

QWhat are the three potential 'singularity' scenarios McCrory discusses that could disrupt the current optimistic view of AI and employment?

AThe three scenarios are: 1) Software Singularity (AI automating the process of innovation itself), 2) Economic Singularity (the economic consequence of software singularity, potentially leading to 'infinite growth in finite time'), and 3) Coasian Singularity (AI agents reducing transaction costs so much that the rationale for firms is undermined).

QWhat is the timeframe Peter McCrory gives for the validity of his judgment that AI will not significantly increase U.S. unemployment?

AHe gives his judgment an 'expiration date' of 12 months. He predicts that one year later, the U.S. unemployment rate will not have risen significantly due to AI.

Lecturas Relacionadas

No Spot Price, How to Trade Anthropic? Entropy Wants to Use Order Books to Price Private Assets

**Summary:** Entropy, operating as a HIP-3 deployer on Hyperliquid, is introducing perpetual contracts for private pre-IPO companies—starting with Anthropic—where no continuous public market price exists. To address this, Entropy employs a novel "liquidity-weighted oracle" that blends its own order book's mid-price (based on a 5-minute EMA) with aggregated external private market data. The oracle’s reliance on internal market prices scales with the "executable depth" within a 200bps range: when order book liquidity is deep and credible, Entropy’s market can command up to 95% weight in pricing; when liquidity thins, the oracle falls back on stale private data. This design aims to prevent manipulation via spoofing and allows the market to generate a continuous consensus price for illiquid assets. If Anthropic remains private at expiry, the ANTH contract will settle based on a 30-day TWAP of its own mark price—essentially settling on the market’s own derived valuation rather than a fundamental or IPO price. Entropy’s bet is that sufficient, credible liquidity can enable real-time price discovery for private assets, closing the gap between infrequent private funding rounds and public market dynamics.

marsbitHace 1 min(s)

No Spot Price, How to Trade Anthropic? Entropy Wants to Use Order Books to Price Private Assets

marsbitHace 1 min(s)

JPMorgan Research Report Analysis: NVIDIA's October Outlook Bullish at $107 Billion, H200 Shipments to China a Variable

JP Morgan's preview report for Nvidia's Q2 FY2027 reiterates an Overweight rating and a $280 price target. While the upcoming quarterly results are expected to beat consensus again, with Q3 guidance potentially reaching $107-108 billion (vs. $104.5B consensus), the bank notes this alone may not drive the stock higher. Historical data shows shares have typically fallen after strong guidance beats. The market now focuses on four key issues: 1) **China H200 shipments**, which have seen preliminary signs and could provide a significant revenue uplift (~$3B per 100k units). 2) **HBM supply constraints**, as evidenced by scaled-back memory configurations for the upcoming Rubin platform. 3) **Competition from AI ASICs/XPUs**, where Nvidia can highlight its platform flexibility and the ~90% per-token cost reduction expected from Vera Rubin vs. Blackwell Ultra. 4) The sustainability of **AI infrastructure investment**. JP Morgan argues Nvidia's valuation is at historically low levels (approx. 17x CY2027 consensus EPS), providing a buffer. The stock's performance hinges on management's ability to address these structural concerns, not just on surpassing numerical forecasts.

marsbitHace 41 min(s)

JPMorgan Research Report Analysis: NVIDIA's October Outlook Bullish at $107 Billion, H200 Shipments to China a Variable

marsbitHace 41 min(s)

Bessent's True Agenda

U.S. Treasury Secretary Besant's recent interventions in the bond market are seen by some as an attempt to artificially trigger a massive short squeeze before the midterm elections. The alleged goal is to push the 10-year yield toward 5% initially, then use covering by short sellers to drive it down to around 4.3%, providing political breathing room for the Trump administration. Analysts point to near-record short positions held by CTA trend-following funds. According to Goldman Sachs data, a price rally of two standard deviations could trigger historically large short-covering. Besant's tactics reportedly include bond buybacks, shifting issuance toward shorter-term debt, and potentially canceling ultra-long bond auctions to pressure these market positions. So far, interventions like modest buybacks have had limited effect on yields, with the Treasury later hinting at using its General Account (TGA) cash for support. Critics argue these measures cannot address structural pressures like large deficits and high inflation. Instead, the strategy appears tactical—aimed at buying time and exploiting technical market vulnerabilities rather than reversing the underlying yield trend. The political timeline is clear: achieve lower yields ahead of the midterms to help lower mortgage rates and craft a favorable narrative. However, analysts warn that once the election passes, structural upward pressure on yields and equity valuations could reassert themselves more forcefully. Market signals suggest this potential short-squeeze dynamic could intensify in the coming days.

marsbitHace 51 min(s)

marsbitHace 51 min(s)

Jackson Hole Speech Preview: Is the Fed Seeking Justification for a Rate Hike?

"Jackson Hole Speech Preview: Is the Fed Looking for Reasons to Raise Rates?" Ahead of the key Jackson Hole symposium, Federal Reserve officials and the IMF are signaling heightened concern over persistent inflation, with some advocating for tighter policy. Boston Fed President Susan Collins, in a recent article, supported holding rates steady only if evidence of falling inflation continues. Otherwise, she argues for prompt tightening, noting inflation has been above target for over five years and warning that prolonged deviation could entrench consumer expectations. Although not a voting member this year, her stance aligns with several officials; three FOMC voters dissented in July, favoring a rate hike. Richmond Fed's Tom Barkin warned of a future "reckoning" regarding the $40 trillion public debt, while IMF Managing Director Kristalina Georgieva urged central banks to maintain a laser focus on price stability. The core dilemma lies in the sources of inflation: Trump-era tariffs, Middle East conflict-driven oil prices, and surging AI investment. Collins believes the first two factors are fading, but identifies AI infrastructure spending as exerting "upward pressure" on core goods inflation. The problem is that interest rate hikes primarily curb demand, not these supply-side shocks. Georgieva framed it as a "tug of war" between negative supply shocks from the Middle East and positive demand shocks from AI. Meanwhile, economic data shows strain. The August Consumer Confidence Index fell to a seven-month low of 89.4. While the present situation index improved, future expectations plummeted. Consumers expect higher inflation (5.8% vs. 5.6% last month), likely influenced by high gas prices. Other data points are weak: July retail sales saw the largest drop in over a year, and job growth stalled. Key upcoming events include the July PCE inflation data (expected to remain well above the 2% target) and new Fed Chair Kevin Warsh's first major speech at Jackson Hole. Market pricing remains conflicted, showing a high probability of a December hike but expecting no move in September. Gold prices, however, have surged over 7% to near three-month highs, signaling market anxiety.

marsbitHace 1 hora(s)

Jackson Hole Speech Preview: Is the Fed Seeking Justification for a Rate Hike?

marsbitHace 1 hora(s)

Today, HYPE Activates Its Second Buyback Engine

"Hyperliquid Activates Second Buyback Engine with AQAv2 Launch" On August 26th, Hyperliquid officially activated its Aligned Quote Assets v2 (AQAv2) mechanism, introducing a major new source of revenue and buyback power for its native token, HYPE. Unlike the existing revenue stream from trading fees, AQAv2 generates income by sharing approximately 90% of the adjusted reserve earnings from stablecoins (primarily USDC) circulating on the Hyperliquid chain. This income is settled every 30 days and automatically funneled into the Assistance Fund for HYPE buybacks, with the first expected distribution on October 3rd. Through a partnership with Circle (issuer of USDC) and Coinbase (treasury manager), Hyperliquid leverages its user base and financial infrastructure to share in the yield generated by stablecoin reserves. This transforms its liquidity and distribution capabilities into a scalable business model. Initial estimates project AQAv2 could generate $150-200 million in annual buyback funds, based on the current ~$6.43 billion USDC supply on Hyperliquid and an assumed reserve yield. Crucially, this revenue is weakly correlated with trading activity and grows directly with the platform's stablecoin adoption. With HYPE recently reaching an all-time high, AQAv2 provides a new fundamental support layer. It shifts Hyperliquid's value capture beyond just transaction fees, anchoring HYPE's long-term buyback potential to the expanding scale of its stablecoin ecosystem.

marsbitHace 1 hora(s)

Today, HYPE Activates Its Second Buyback Engine

marsbitHace 1 hora(s)

Trading

Spot