#laboreconomics — Public Fediverse posts

Fear and Loathing of AI (Part III): “Learn AI” Is the New “Learn to Code”

By Cliff Potts, CSO, and Editor-in-Chief of WPS News

There is a sentence that shows up in every technological cycle right before the disappointment phase begins.

“Just learn the skill.”

It sounds empowering. It sounds reasonable. It sounds like personal agency.

It is also a lie we have been telling people for decades.

The obedience script

“Learn to code” was never about opportunity.
It was about discipline.

It trained people to accept that:

• structural failures are personal problems,
• economic insecurity is an individual moral test,
• and survival depends on constant retraining at your own expense.

When the promised jobs didn’t materialize—or paid far less than advertised—the story shifted seamlessly: you didn’t learn the right language, the right framework, the right stack.

Now the phrase has been updated.

“Learn AI.”

Same script. Same pressure. Same outcome.

Skills don’t collapse — markets do

Coding did not fail because people were lazy or incapable. It failed because markets flooded, tools commoditized, and labor lost leverage.

AI will follow the same arc, only faster.

The moment a skill becomes:

• widely accessible,
• easily automated,
• and expected rather than rewarded,

it stops being a path to security and becomes a baseline requirement for staying afloat.

The reward for compliance is not prosperity.
It is continued participation.

Training as cost transfer

Here is what “learn AI” really means in practice:

• You pay for the courses.
• You absorb the time cost.
• You shoulder the career risk.
• You adapt repeatedly as tools change.
• You accept lower pay because “AI makes you more efficient.”

None of that is accidental.

It is a system designed to push costs downward while extracting value upward.

The more often you are told to retrain, the clearer it becomes that training itself is the product.

The illusion of agency

People are encouraged to believe that mastery equals control.

But control does not come from skill alone.
It comes from:

• ownership,
• bargaining power,
• regulation,
• and collective leverage.

Without those, skill is just labor dressed up as self-improvement.

Learning AI may help you keep your job a little longer.
It will not protect you from the logic of the system deploying it.

What learning actually means now

This does not mean you should refuse to learn.

It means you should learn without illusions.

Learn AI the way you learn any tool:

• to reduce friction,
• to save time,
• to extend what you already do.

Do not learn it expecting salvation.
Do not learn it expecting loyalty from platforms.
Do not learn it expecting the market to reward you for effort.

Markets reward leverage, not diligence.

The quiet truth

The most dangerous part of “learn AI” is not that it is false.

It is that it is incomplete.

It tells people how to adapt, but never who benefits.
It demands flexibility, but never offers stability.
It promises relevance, but never guarantees dignity.

We have seen this cycle before.

And it did not end with freedom.

It ended with exhaustion.

For more social commentary, please see Occupy 2.5 at https://Occupy25.com

TVs get cheaper every year. So does fast fashion. But rent, healthcare, and college tuition keep climbing. Baumol explained this in 1966. We're still living it. by @daylightatheism.bsky.social

https://onlys.ky/cost-disease/

#Economics #CostOfLiving #Inequality #LaborEconomics #PublicPolicy

Choice Architecture in Occupational Choices
http://repec.business.uzh.ch/RePEc/iso/leadinghouse/0255_lhwpaper.pdf
This study uses a Swiss job board to analyze how rank order and design influence high-stakes occupational choices. Higher rankings increased applications, especially for high-paying and gender-congruent occupations. Users interpreted rank to justify choices aligning with identity, providing field evidence for motivated reasoning. An interactive, visually enriched interface redesign boosted applications and watch list usage. Results show that reducing cognitive load expands the variety of options individuals consider and remember.
#choicearchitecture #motivatedreasoning #laborEconomics #jobtech #ExperimentalEcon
#BoundedRationality

Extending Working Lives: A Systematic Review of Motivations, Determinants, and Institutional Contexts
https://cris.maastrichtuniversity.nl/ws/files/307829408/ROA_TR_2026_2_Extending_Working_Lives.pdf
This review of 103 studies examines determinants of labor participation beyond #retirement age across diverse institutional contexts. Causal evidence indicates that pension reforms and tax incentives yield only modest impacts on extending working lives. Instead, employer practices and workplace flexibility act as primary determinants for feasible post-retirement employment. Findings reveal that financial necessity drives liberal systems, whereas intrinsic motives characterize social-democratic contexts. Success requires aligning macro-level policy with firm-level adaptations to support older workers.

…with generous references to
Working beyond retirement age in Germany: The employee’s perspective https://economicscience.net/publications/2011-les/
…individuals with lower income or wealth are more likely to continue working past the statutory retirement age
…financial security enables full labor market withdrawal
… functional capacity predicts the ability to continue working but not necessarily the intention to do so
… work motivation predicts willingness to remain employed beyond retirement age in Germany, partly through its positive association with self-reported work ability and openness to further education. Job rewards also increase willingness to continue working.

#activeaging #pensionreform #laborEconomics

Do Firms Share their Profits Equally with Women and Men? The Role of Human Capital, Managerial Positions and Unions https://docs.iza.org/dp18388.pdf
"… wage-profit elasticity is estimated at 2.8% and is not statistically different for women and men. These non-differing elasticities therefore imply a non-significant price effect in the gender wage gap, which is estimated in our analysis at 15.6%
… higher human capital – measured here by education level or tenure – and holding a managerial position increase rent-sharing for both men and women
… Still, rent-sharing seems to fuel the gender wage gap, albeit to a fairly modest extent (at around 5% of the gender wage gap for our benchmark specification) through the channel of segregation (i.e. women are somewhat more concentrated in less profitable firms)"
#gpg #wages #LaborEconomics

Do Firms Share their Profits Equally with Women and Men? The Role of Human Capital, Managerial Positions and Unions https://docs.iza.org/dp18388.pdf
"… wage-profit elasticity is estimated at 2.8% and is not statistically different for women and men. These non-differing elasticities therefore imply a non-significant price effect in the gender wage gap, which is estimated in our analysis at 15.6%
… higher human capital – measured here by education level or tenure – and holding a managerial position increase rent-sharing for both men and women
… Still, rent-sharing seems to fuel the gender wage gap, albeit to a fairly modest extent (at around 5% of the gender wage gap for our benchmark specification) through the channel of segregation (i.e. women are somewhat more concentrated in less profitable firms)"
#gpg #wages #LaborEconomics

Do Firms Share their Profits Equally with Women and Men? The Role of Human Capital, Managerial Positions and Unions https://docs.iza.org/dp18388.pdf
"… wage-profit elasticity is estimated at 2.8% and is not statistically different for women and men. These non-differing elasticities therefore imply a non-significant price effect in the gender wage gap, which is estimated in our analysis at 15.6%
… higher human capital – measured here by education level or tenure – and holding a managerial position increase rent-sharing for both men and women
… Still, rent-sharing seems to fuel the gender wage gap, albeit to a fairly modest extent (at around 5% of the gender wage gap for our benchmark specification) through the channel of segregation (i.e. women are somewhat more concentrated in less profitable firms)"
#gpg #wages #LaborEconomics

Do Firms Share their Profits Equally with Women and Men? The Role of Human Capital, Managerial Positions and Unions https://docs.iza.org/dp18388.pdf
"… wage-profit elasticity is estimated at 2.8% and is not statistically different for women and men. These non-differing elasticities therefore imply a non-significant price effect in the gender wage gap, which is estimated in our analysis at 15.6%
… higher human capital – measured here by education level or tenure – and holding a managerial position increase rent-sharing for both men and women
… Still, rent-sharing seems to fuel the gender wage gap, albeit to a fairly modest extent (at around 5% of the gender wage gap for our benchmark specification) through the channel of segregation (i.e. women are somewhat more concentrated in less profitable firms)"
#gpg #wages #LaborEconomics

Do Firms Share their Profits Equally with Women and Men? The Role of Human Capital, Managerial Positions and Unions https://docs.iza.org/dp18388.pdf
"… wage-profit elasticity is estimated at 2.8% and is not statistically different for women and men. These non-differing elasticities therefore imply a non-significant price effect in the gender wage gap, which is estimated in our analysis at 15.6%
… higher human capital – measured here by education level or tenure – and holding a managerial position increase rent-sharing for both men and women
… Still, rent-sharing seems to fuel the gender wage gap, albeit to a fairly modest extent (at around 5% of the gender wage gap for our benchmark specification) through the channel of segregation (i.e. women are somewhat more concentrated in less profitable firms)"
#gpg #wages #LaborEconomics

Turns Out They Didn’t Really Want You To Bring Your Whole Self To Work

https://fed.brid.gy/r/https://www.techdirt.com/2026/02/03/turns-out-they-didnt-really-want-you-to-bring-your-whole-self-to-work/

Interesting angle, but letting your best people walk just to avoid paying them more seems like self-sabotage. Sure, you keep “solid” workers at standard wages, but isn’t that how you end up with mediocrity baked in? #WorkCulture #HR #BusinessStrategy #LaborEconomics

Why top firms paradoxically fi...

Pluralistic: The enshittification of labor (07 Nov 2025)

https://web.brid.gy/r/https://pluralistic.net/2025/11/07/postwar-social-contract/

Things versus People: Gender Differences in Vocational Interests and in Occupational Preferences https://www.sciencedirect.com/science/article/pii/S0167268122003201
HT @SteveStuWill
"On average, men are more interested in working with things, whereas women are more interested in working with people.
…
Every push to get more women into things-related fields is simultaneously a push to get them out of equally important people-related fields such as healthcare and education - fields that, on average, they tend to prefer. We’re therefore faced with a decision: Should #GenderEquality mean identical outcomes, or identical freedom to follow one’s interests?"
#STEM #LaborEconomics #Psychology

AI-assisted Programming May Decrease the Productivity of Experienced Developers by Increasing Maintenance Burden https://arxiv.org/pdf/2510.10165
"… find that productivity indeed increases.
… the increase in productivity is driven by less-experienced (peripheral) developers.
… find that code written after the adoption of #AI requires more rework.
… the added rework burden falls on the more experienced (core) developers, who review 6.5% more code after Copilot's introduction, but show a 19% drop in their original code #productivity.
… this finding raises caution that productivity gains of AI may mask the growing burden of maintenance on a shrinking pool of experts."
#economics #LaborEconomics #skills

Thrilled to share my latest publication with coauthors Marianne Tønnessen and Astri Syse (open access) in Demography, October 2025. I thank the editor Sara Curran, and three anonymous reviewers, and #ReadDemography, and #PennPSC. https://read.dukeupress.edu/demography/article/doi/10.1215/00703370-12271149/403699/Goodbye-Norway-Top-Earners-Selection-Into-Return?fbclid=IwY2xjawNfgvtleHRuA2FlbQIxMQABHgkapuzYnl7R8qkJqdJofVHsHkPvm8I2VVzefb53l-ezAkYI_o64mtK7vNMf_aem_bc5nosvA7cc3zdoRsO4vuQ #Migration, #LaborEconomics, #PopulationEconomics, #norway

Mind the Gap: Gender-based Differences in Occupational Embeddings
https://aclanthology.org/2025.gebnlp-1.7.pdf
"Across five state-of-the-art multilingual models and seven reference-set configurations, up to 82% of gendered pairs received divergent Top-5 suggestions. These differences involved distinct occupational codes that sometimes crossed major #KldB group

.…gendered job titles—such as Autor vs. Autorin —often lead to different occupation codes, despite having identical meanings. Our findings underscore the importance of grounding #NLP innovations in language-specific sociolinguistic knowledge. Without rigorous attention to linguistic structure and social context, these tools risk perpetuating systemic biases—particularly in settings where semantic equivalence is masked by morphological variation. Addressing such challenges is crucial not only for the technical refinement of NLP systems, but for ensuring that their real-world applications advance rather than hinder equity"
#jobtech #gender #discrimination #LaborEconomics #llm

Measuring Gender Bias in Job Title Matching for Grammatical Gender Languages
https://www.arxiv.org/pdf/2509.13803
"… propose a methodology to measure gender bias in a high-impact #NLP application in the human resources domain: job title matching. Using an existing test set in English for this task, we have generated gender-annotated analogous corpora in four languages with grammatical gender, and addressed the evaluation of #genderBias as ranking comparison controlling for gender. Additionally, we establish baselines and confirm that this type of bias already exists in out-of-the-box pre-trained models, which are often used as the core for developing job title matching applications.

Finding a trade-off between model performance and #gender #bias is an important issue to address when developing and selecting job matching models for deployment. On the one hand, choosing a model with apparent good performance but that in turn shows a considerable gender gap may not only be ethically questionable, but it may also result in reputation and even legal consequences on the company responsible for it."
#llm #jobtech #discrimination #LaborEconomics

Against the Standard https://archiv.ub.uni-heidelberg.de/volltextserver/36879/8/Cubel_Against_dp764_2025.pdf
"… in the absence of feedback, women are less likely than men to benchmark their performance against a standard of excellence. This is inefficient because women who are likely to obtain increased rewards choose a low reward scheme instead.
… When feedback is provided and the standard is set by peers, this gender gap closes. However, the gap re-emerges, and even widens, when the standard of excellence is set by experts.
… If standards are set by experts and committees are perceived as male-dominated, a gender gap will exist in the award of promotions, grants or recognition. Understanding the differential impact of standards and feedback provision can help to design more inclusive competitive processes and bridge gender gaps in labour market outcomes."
#ExperimentalEcon #LaborEconomics #wages #gpg #LaborMarkets

A job-based assessment of economic complexity: from hidden to revealed https://arxiv.org/pdf/2507.05846
"… develop a multipartite framework to connect four layers of nodes: #skills, #occupations, #industries, and US counties. From the skill-occupation network, we algorithmically compute an assessment of the economic #complexity, or fitness, of jobs. Then we compute our estimate of the complexity of industries by averaging the fitness values of the jobs active in a specific industry. Finally, the fitness of US counties is the sum of the complexity of the industries that operate in that territory.
… approach has a number of advantages. First, it allows for a natural inclusion of services, a fundamental aspect that is often neglected in the economic complexity investigations because of data unavailability. Second, it permits the computation of the fitness of those counties with a small or null manufacturing activity. Third, it solves the numerical problems of the algorithmic assessments of the economic complexity indicators, namely the presence of very small values and the emergence of multimodal distributions. Even more importantly, the hidden complexity is positively associated with local economic #growth and predicts both wage levels and labor #productivity growth"
#wages #LaborEconomics

Large Language Models, Small Labor Market Effects
https://bfi.uchicago.edu/wp-content/uploads/2025/04/BFI_WP_2025-56-2.pdf
"#AI chatbots have had no significant impact on earnings or recorded hours in any occupation, with confidence intervals ruling out effects larger than 1%. Modest productivity gains (average time savings of 3%), combined with weak wage pass-through, help explain these limited labor market effects.
… no evidence of differential trends over time, suggesting that the limited effects are not merely a very short-run phenomenon.
… findings challenge narratives of imminent labor market transformation due to Generative AI."
#llm #laborEconomics #wages

Remote Work, Employee Mix, and Performance
https://cevatgirayaksoy.wordpress.com/wp-content/uploads/2025/06/w33851.pdf
"…fully remote work increased the share of women, including married women, rural and smaller-town residents. By accessing groups with traditionally lower labor-force participation, the firm was able to increase its share of graduate employees by 14% without raising #wages
…workforce productivity rose by 10%, reflecting shorter call durations for remote employees. This was facilitated by a quieter home working environment, avoiding the background noise in the office
…fully remote employees with initial in-person training saw higher long-run remote #productivity and lower #attrition rates."
#wfh #laborEconomics

A recent study analyzing the Danish labor market (2023-2024) offers some insights into the early impact of generative AI. While AI chatbots have seen rapid adoption across various occupations, the study suggests their effects on overall wages and employment have been remarkably limited so far. 🌐

Key takeaways from the research:
📈 Despite widespread use, AI chatbots showed no significant impact on earnings or recorded hours.
🤔 While AI saves time for many users, the study found that new tasks created by AI (like reviewing AI output or crafting prompts) often offset these time savings for 8.4% of workers.
💸 Only a small fraction (3-7%) of productivity gains translated into higher earnings for workers, raising questions about who benefits most from efficiency.

This early look provides perspective, challenging some narratives around immediate, widespread labor market transformation. It highlights that the integration of AI is still evolving and its long-term economic impact remains a dynamic area for further research. What are your observations on AI's impact in your workplace?
https://arstechnica.com/ai/2025/05/time-saved-by-ai-offset-by-new-work-created-study-suggests/
#GenerativeAI #LaborEconomics #AIImpact #WorkplaceInnovation #FutureOfWork

Behavioral Measures Improve AI Hiring: A Field Experiment https://d.repec.org/n?u=RePEc:rco:dpaper:532&r=&r=ain
"… suggest that survey-based behavioral measures markedly improve the predictions of a random-forest algorithm trained to predict productivity within sample relative to demographic information alone."

It's a pity that the authors do not give the more traditional probit model as much attention as their fancy "#AI", a random forrest model. They spend a lot of effort to find a good random forrest model with cross validation. But it is pitted against a simple probit model where they didn't even try to include interaction effects according to their description. Now, what is the computational cost of the cross validated random forrest model compared to a well crafted probit model? Of course you can do automated feature and feature interteraction selection with probit models, too. There is no reason to dismiss the probit model in such an unfair comparison.
#jobtech #LaborEconomics #ML

Who Gets the Callback? Generative AI and Gender Bias https://d.repec.org/n?u=RePEc:arx:papers:2504.21400&r=&r=lma
"… most #llm models reproduce stereotypical gender associations and systematically recommend equally qualified women for lower-wage roles, indicating occupational segregation.
… These biases stem from entrenched gender patterns in the training data as well as from an agreeableness bias induced during the reinforcement learning from human feedback stage
.…AI-driven hiring may perpetuate biases in the labor market and have implications for #fairness and diversity within firms"
#AI #jobtech #ExperimentalEcon #LaborEconomics #discrimination #bias

Time saved by #AI offset by new work created https://arstechnica.com/ai/2025/05/time-saved-by-ai-offset-by-new-work-created-study-suggests/
"Despite finding widespread and often employer-encouraged adoption of these tools, the study concluded that “AI chatbots have had no significant impact on earnings or recorded hours in any occupation” during the period studied. The confidence intervals in their statistical analysis ruled out average effects larger than 1%."
https://papers.ssrn.com/sol3/Delivery.cfm/5219933.pdf?abstractid=5219933&mirid=1
#economics #LaborEconomics

#valueoflabor #workerworth #wagereality #productivitygap #laboreconomics #businessbasics #economicsexplained #careerinsights #workplacewisdom #moneyfacts #surplusvalue #wagesuppression #modernexploitation #yourlabormatters #whocreatesprofits #visualizedata #everydaymath #knowyourworth #worklifebalance #mastodoneconomics

Half-Past Four is the New Five O’Clock in More Efficient Workday
https://archive.ph/PNdyP
The average American workday now ends 42 minutes earlier than two years ago. Despite the shorter workday, overall productivity has increased by about 2%.
#laborEconomics #work

Macroeconomic Impact of Artificial Intelligence on Productivity: An estimate from a survey https://d.repec.org/n?u=RePEc:eti:dpaper:24084&r=&r=ain
"… highly educated and high-wage workers are more likely to use #AI
… the diffusion of AI may widen overall labor market inequality
… estimate that macroeconomic productivity impact is 0.5–0.6% when AI is used than when it is not.
… as approximately 28% of the respondents expect to use AI for their jobs in the future, the macroeconomic effects of AI are likely to expand. However, because the productivity gain of AI for those who have recently started using AI is smaller than that for those who have been using AI continuously, the additional productivity gain is likely to diminish over time."
#economics #LaborEconomics

Estimating Wage Disparities Using Foundation Model https://www.gsb.stanford.edu/gsb-box/route-download/625707
"… classic problem from #laborEconomics: estimating how individuals with the same labor market experience get paid when they belong to different groups. We highlighted the promise of using a foundation model in this setting: wage predictions improve over econometric baselines by 15%. We also showed that an omitted variable bias arises when a foundation model discards relevant information about group differences."
#llm

via @ckrafftc
Research in Labor Economics (RLE) is planning a volume highlighting research on New Developments in Labor Economics. The editors Solomon Polachek and @ben_elsner are soliciting up to ten new papers showcasing new developments in #laborEconomics.
https://legacy.iza.org/rle_application/formApplication

New research reveals that longer #NoticePeriods help workers find better jobs and reduce #unemployment spells. Thus long-term production gains outweigh short-term #Productivity losses. In contrast, larger severance delays job finding and has no impact on wages. #LaborMarket #LaborEconomics

https://academic.oup.com/qje/advance-article/doi/10.1093/qje/qjae029/7810272

New research shows that #inflation affects workers' #wellbeing beyond its impact on real #wages . As prices rise, #workers must fight for raises, creating #ConflictCosts that more than double inflation's effect on their well-being. #LaborEconomics #FutureOfWork
https://www.nber.org/papers/w32956

Unlocking NACE Classification Embeddings with OpenAI for Enhanced Analysis and Processing https://arxiv.org/pdf/2409.11524
"…proposes a methodology for transforming a qualitative classification, such as the NACE classification, into a numerical representation while retaining the original structure, along with methods to evaluate the efficacy of such a transformation."
#llm #LaborEconomics

#CfP
How is #EUIntegration reshaping work in #SoutheasternEurope? We're looking for papers on #labor market reforms, skills & education, labor mobility and technology transfer for a special issue of our open access journal #COMPSEES.
🗓️ Deadline: July 1, 2024
➡️ https://leibniz-ios.de/en/knowledge-transfer/news/newsdetail/cfp-compsees-sonderausgabe
#Economics #LaborEconomics #Histodons #SocialScience #SocialSciences

From Helping Hand to Stumbling Block: The ChatGPT Paradox in #Competency Experiment http://d.repec.org/n?u=RePEc:ces:ceswps:_11002&r=ain
ChatGPT improved reading & writing performance of participants with intermediate skills while math score decreased
…low-ability subjects couldn’t asses the quality of answers
…if one has already adequate skills, the benefits of using #AI may be negligible
…contests idea that AI can complement workers with less #expertise
#LaborEconomics

Generative #AI Usage and #AcademicPerformance https://arxiv.org/pdf/2404.19699
"… students using GenAI tools score on average 6.71 (out of 100) points lower than non-users. While GenAI tools may offer benefits for #learning and engagement, the way students actually use it correlates with diminished academic outcomes.
… ascribe this result primarily to a learning-hindering mechanism when using GenAI"
#LaborEconomics

Too old to be a diversity hire: choice bundling shown to increase gender-diverse hiring decisions fails to increase age diversity
http://eprints.lse.ac.uk/120910/3/Jolles_too_old_to_be_a_diversity_hire_published.pdf
"…found evidence of bias against older job candidates in hiring decisions but found inconsistent effects of choice bundling on the selection of older candidates across experiments."
#ageDiscrimination #LaborEconomics #ExperimentalEcon

Pay more

Unemployed Gen Zers are having to turn work down because they can’t afford the commute and uniform, report shows | Fortune
https://fortune.com/europe/2024/02/05/unemployed-gen-z-turn-work-down-because-they-cant-afford-commute-uniform-report-cost-of-living-crisis/ #wages #labor #laboreconomics #economy #wagegap #affordability #affordabilitycrisis

Nothing Really Matters: Evaluating Demand-Side Moderators of Age #Discrimination in Hiring http://d.repec.org/n?u=RePEc:iza:izadps:dp16672&r=exp
"… senior candidates experience discrimination during the hiring process. On average, they receive 16.97% fewer positive responses than comparable younger candidates
… none of the investigated demand-side characteristics moderated the #ageDiscrimination observed in the field experiment"
#LaborMarkets #LaborEconomics #ExperimentalEcon

External pay transparency and the gender wage gap http://d.repec.org/n?u=RePEc:zbw:rwirep:1027&r=lma
"… providing publicly available wage information in vacancies, so-called external #payTransparency, can reduce the #GenderWageGap… reduction in the gender wage gap was caused by an increase in women's earnings, particularly at the lower part of the distribution. Earnings of men, on the other side, remained largely constant"
#wages #LaborEconomics

More hashtags for better discovery:
#LaborEcon #LaborEconomics #IndustrialRelations #Wages #work #ExperimentalEcon #ExperimentalEconomics #BehavioralEcononomics #CognitiveEconomics #EconomicPsychology #IndustrialOrganization #GameTheory #Microeconomics #Econometrics #SocialNorms #SocialPreferences #Reciprocity #BoundedRationality #heterogeneity

#Erdős4

#Typography #PlainText #Linux #cli #Accessibility #MultipleSclerosis