pith. sign in

arxiv: 2605.00843 · v1 · submitted 2026-04-07 · 💻 cs.CY · cs.AI

Generative-AI and the transformation of workforce. A job postings-driven analysis

Diana Maria Popa , Simona-Vasilica Oprea , Adela B\^ara This is my paper

Pith reviewed 2026-05-10 18:23 UTC · model grok-4.3

classification 💻 cs.CY cs.AI
keywords generative AIjob postingsworkforce skillsskill transformationhybrid expertiselabor marketAI adoption
0
0 comments X

The pith

Job postings show generative AI driving a sharp rise in hybrid human-AI skills since 2021.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper analyzes more than 150,000 English-language job postings spanning 2018 to 2025 to map how generative AI changes the skills employers list as essential. It documents a clear post-2021 jump in requests for abilities such as prompt engineering, fine-tuning, and model validation, while mentions of routine work like data entry and manual coding fall. The authors also track whether postings present AI as something that helps or replaces people and project continued growth in AI-related and soft skills through 2025. If these patterns hold, the labor market is moving toward roles that require workers to combine human judgment with AI tools rather than pure automation.

Core claim

By compiling and processing a multi-source corpus of job postings with lexical extraction, topic modeling, and time-series methods, the study establishes a sharp post-2021 increase in AI skill mentions such as prompt engineering, fine-tuning, and model validation alongside a decline in routine tasks such as data entry and manual coding. Forecasts from the data indicate sustained expansion of AI_Data and Soft_Meta skills through 2025, which the authors interpret as evidence of a structural convergence on hybrid human-AI expertise as the emerging foundation of employability.

What carries the argument

A Framing Index computed from sentence-transformer embeddings and cosine similarity on job-posting text, used together with lexical skill extraction and ARIMA forecasting to separate augmentative from substitutive language across time and sectors.

If this is right

  • Demand for prompt engineering, model validation, and related AI interaction skills will continue to rise in most sectors.
  • Routine data-entry and basic coding positions will keep declining as employers expect AI tools to handle those tasks.
  • Workers who combine domain knowledge with the ability to direct and verify AI outputs will gain a lasting employability advantage.
  • Cross-sector analyses will show increasing interdependence between AI_Data skills and Soft_Meta competencies such as adaptability.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Training programs could prioritize teaching people how to frame effective prompts and evaluate AI outputs alongside traditional domain content.
  • Labor-market monitoring agencies might adopt similar large-scale posting analysis to detect skill shifts earlier than official statistics allow.
  • Role design inside firms may increasingly split tasks between AI execution and human oversight, changing job descriptions across industries.

Load-bearing premise

That the wording in public job postings reliably tracks what employers actually need and will hire for, rather than reflecting only recruitment language or selection effects.

What would settle it

A follow-up survey of recent hires or employer interviews showing that the actual skills required on the job have not shifted in the same directions and magnitudes indicated by the postings trends.

Figures

Figures reproduced from arXiv: 2605.00843 by Adela B\^ara, Diana Maria Popa, Simona-Vasilica Oprea.

Figure 1
Figure 1. Figure 1: Full analytical pipeline 3.2.1. Text preprocessing and structuring The analytical framework is implemented as a modular pipeline designed to transform the heterogeneous corpus of job descriptions into a structured, analyzable format suitable for semantic modeling, skill extraction and temporal analysis. The records in the final dataset contain two essential variables: a publication date, ranging from 2018 … view at source ↗
Figure 2
Figure 2. Figure 2: Skill mentions per 1000 job postings 4.2. Semantic-shift and framing analysis As shown in the left plot in [PITH_FULL_IMAGE:figures/full_fig_p012_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Temporal dynamics of AI-related semantic proximity and framing index (2018-2025) 4.3. Sectoral analysis The exceptionally high concentration of domain-specific terminology in Healthcare indicates that automation and AI systems have not yet penetrated deeply into clinical or diagnostic roles (as in [PITH_FULL_IMAGE:figures/full_fig_p012_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Evolution of AI mentions across sectors and corresponding skill profiles [PITH_FULL_IMAGE:figures/full_fig_p013_4.png] view at source ↗
Figure 8
Figure 8. Figure 8: Wordclouds-LDA In the third topic, words such as consultant, client, marketing, media and team reveal an orientation toward digital marketing, creative consultancy and AI-assisted communication strategies. This cluster highlights how AI tools are embedded in marketing ecosystems, reshaping recruitment, branding and customer engagement through personalized and data-driven media practices. The fourth topic, … view at source ↗
Figure 9
Figure 9. Figure 9: Wordclouds LDA 4.5.4. Comparison of the results A comparison of the results obtained by the three techniques is offered in [PITH_FULL_IMAGE:figures/full_fig_p017_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Semantic framing across sectors Findings showed a consistent trend toward augmentation-oriented framing, especially after 2021, when the adoption of generative-AI technologies accelerated. In e-commerce, this tendency was particularly pronounced, with frequent references such as AI-assisted marketing, decision-support analytics, and human-in-the-loop optimization. These expressions framed AI as a co-creat… view at source ↗
Figure 11
Figure 11. Figure 11: Forecast The oscillatory ARIMA(2,0,2) model captures cyclical fluctuations that the smoothed specification attenuates. Notably, Soft_Meta and Leadership continue to expand moderately, but with visible short-term oscillations, indicating alternating phases of emphasis on adaptive and managerial capabilities. Domain_Specific remains robust yet shows slight periodic corrections, which may reflect the saturat… view at source ↗
read the original abstract

This paper investigates how generative-artificial intelligence AI is reshaping job requirements, skill compositions and sectoral dynamics across global labor markets. It examines the evolving frequency and framing of AI-related competencies in job postings, exploring whether generative-AI functions primarily as an augmentative or substitutive force in the workplace. A large-scale, multi-source corpus of over 150,000 English-language job postings 2018-2025 is compiled from twelve open-access datasets and one public API. The analytical framework integrates lexical skill extraction, semantic framing, topic modeling, BERTopic, LDA, KMeans, and time-series forecasting ARIMA. Skill mentions are categorized into five dimensions: AI_Data, Routine, Soft_Meta, Domain_Specific and Leadership, while cross sectoral analyses and correlation matrices quantify interdependencies between competencies. Sentence-transformer embeddings and cosine similarity are used to compute a Framing Index, distinguishing augmentation- versus automation-oriented discourse. Investigating job postings, our research contributes a replicable, data driven methodology for mapping the diffusion of AI related skills across industries and time. Results reveal a sharp post-2021 increase in AI-related skill mentions: prompt engineering, fine-tuning and model validation, accompanied by a decline in routine tasks: data entry and manual coding. Forecasts suggest sustained growth in AI_Data and Soft_Meta skills through 2025, signaling a structural convergence toward hybrid human-AI expertise as a new foundation of employability.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

4 major / 3 minor

Summary. This paper analyzes a corpus of over 150,000 English-language job postings from 2018 to 2025 collected from multiple open-access sources to investigate the transformation of workforce skills due to generative AI. Using lexical skill extraction, topic modeling techniques including BERTopic, LDA, and KMeans, and a Framing Index based on sentence-transformer embeddings and cosine similarity, the authors categorize skills into AI_Data, Routine, Soft_Meta, Domain_Specific, and Leadership dimensions. They report a sharp increase in mentions of AI skills such as prompt engineering and fine-tuning after 2021, a decline in routine tasks like data entry, and use ARIMA models to forecast continued growth in AI-related and soft skills through 2025, suggesting a convergence to hybrid human-AI expertise.

Significance. If the assumptions underlying the skill extraction and framing analysis hold, this study would offer significant empirical insights into how generative AI is reshaping job requirements across sectors. The large-scale, multi-source dataset and the integration of multiple analytical methods provide a replicable framework for tracking AI skill diffusion, which could inform workforce development policies and educational curricula. The forecasting component adds forward-looking value, though its reliability depends on the robustness of the underlying models.

major comments (4)
  1. [Analytical framework] The Framing Index, computed via sentence-transformer embeddings and cosine similarity to distinguish augmentation- versus automation-oriented discourse, is central to the claim about AI's augmentative or substitutive role; however, no human-annotation benchmark, inter-annotator agreement metrics, or robustness checks across embedding models are reported, leaving open the possibility of embedding biases or selection effects in the postings.
  2. [Data collection and preprocessing] The compilation of the 150,000+ job postings from twelve datasets lacks detailed description of cleaning procedures, deduplication, language filtering, and any validation for the lexical skill extraction process, which directly impacts the reliability of the reported frequency changes in AI-related and routine skills post-2021.
  3. [Topic modeling and skill categorization] The categorization of skills into five dimensions (AI_Data, Routine, Soft_Meta, Domain_Specific, Leadership) using LDA, KMeans, and BERTopic involves free parameters such as the number of topics, with no sensitivity analysis or stability checks provided to support the cross-sectoral analyses and correlation matrices.
  4. [Time-series forecasting] The ARIMA forecasts suggesting sustained growth in AI_Data and Soft_Meta skills through 2025 depend on model orders (p, d, q) that are fitted but not accompanied by robustness checks, alternative specifications, or confidence intervals, which are necessary to substantiate the structural convergence claim.
minor comments (3)
  1. [Abstract] The abstract refers to 'twelve open-access datasets and one public API' but does not list them; providing the specific sources would enhance replicability.
  2. [Methods] The notation and exact formula for the Framing Index should be presented as an equation for clarity.
  3. [Results] Trend figures would benefit from including error bars or confidence intervals to reflect uncertainty in the skill mention frequencies.

Simulated Author's Rebuttal

4 responses · 0 unresolved

We thank the referee for the thorough and constructive review of our manuscript. We address each major comment point by point below, providing the strongest honest responses possible based on the current version of the paper. We will incorporate revisions to enhance transparency, robustness, and replicability where the concerns are valid.

read point-by-point responses

  1. Referee: [Analytical framework] The Framing Index, computed via sentence-transformer embeddings and cosine similarity to distinguish augmentation- versus automation-oriented discourse, is central to the claim about AI's augmentative or substitutive role; however, no human-annotation benchmark, inter-annotator agreement metrics, or robustness checks across embedding models are reported, leaving open the possibility of embedding biases or selection effects in the postings.

    Authors: We acknowledge that the current manuscript does not report a human-annotation benchmark, inter-annotator agreement metrics, or robustness checks across embedding models for the Framing Index. This is a valid concern regarding potential biases. In the revised manuscript, we will add a dedicated validation subsection that includes a human annotation study on a random sample of 500 job postings (with two independent annotators), compute Cohen's kappa for inter-annotator agreement, and perform robustness checks using alternative sentence-transformer models (e.g., all-MiniLM-L6-v2 vs. paraphrase-MiniLM-L12-v2). These additions will directly address the possibility of embedding biases and strengthen the central claim. revision: yes

  2. Referee: [Data collection and preprocessing] The compilation of the 150,000+ job postings from twelve datasets lacks detailed description of cleaning procedures, deduplication, language filtering, and any validation for the lexical skill extraction process, which directly impacts the reliability of the reported frequency changes in AI-related and routine skills post-2021.

    Authors: We agree that the Methods section currently provides insufficient detail on data cleaning, deduplication, language filtering, and validation of lexical skill extraction. This limits replicability. We will expand the Data section in the revision to specify: (i) deduplication via Jaccard similarity threshold of 0.85 combined with exact URL matching, (ii) language filtering using langdetect with a 0.9 confidence threshold retaining only English postings, (iii) cleaning steps including removal of HTML artifacts and standardization of date formats, and (iv) validation of lexical extraction via precision/recall evaluation on a manually labeled subset of 1,000 postings. These details will support the reliability of the post-2021 frequency changes reported. revision: yes

  3. Referee: [Topic modeling and skill categorization] The categorization of skills into five dimensions (AI_Data, Routine, Soft_Meta, Domain_Specific, Leadership) using LDA, KMeans, and BERTopic involves free parameters such as the number of topics, with no sensitivity analysis or stability checks provided to support the cross-sectoral analyses and correlation matrices.

    Authors: We recognize that the absence of sensitivity analysis for topic modeling parameters (e.g., number of topics in LDA and BERTopic, cluster count in KMeans) is a limitation that affects confidence in the five-dimension categorization and downstream cross-sectoral results. We will revise the manuscript to include a sensitivity analysis subsection: varying the number of topics from 4 to 15, reporting topic coherence (C_v) and stability across 10 random seeds for each method, and providing silhouette scores for KMeans. This will demonstrate the robustness of the skill categorization and support the correlation matrices and sectoral analyses. revision: yes

  4. Referee: [Time-series forecasting] The ARIMA forecasts suggesting sustained growth in AI_Data and Soft_Meta skills through 2025 depend on model orders (p, d, q) that are fitted but not accompanied by robustness checks, alternative specifications, or confidence intervals, which are necessary to substantiate the structural convergence claim.

    Authors: We accept that the ARIMA forecasting section lacks reported model orders, robustness checks, alternative specifications, and confidence intervals, which weakens the structural convergence claim. In the revision, we will explicitly state the fitted (p, d, q) orders for each skill dimension, include 95% confidence intervals on the forecasts, add robustness checks via alternative models (e.g., SARIMA and Holt-Winters exponential smoothing), and compare forecast accuracy using AIC/BIC and out-of-sample validation on 2023-2024 holdout data. These changes will better substantiate the forecasts of continued growth in AI_Data and Soft_Meta skills. revision: yes

Circularity Check

0 steps flagged

No circularity: empirical pipeline uses external data and standard computations

full rationale

The paper compiles an external corpus of 150k+ job postings from open-access datasets and applies lexical extraction, BERTopic/LDA/KMeans topic modeling, sentence-transformer cosine similarity for the Framing Index, and ARIMA forecasting. Core observations (post-2021 rise in prompt engineering/fine-tuning mentions, decline in routine tasks) are direct frequency counts and correlations from the raw data, not outputs of any fitted model. ARIMA fits are used only for forward extrapolation to 2025 and do not redefine or validate the historical skill frequencies. The Framing Index is a standard embedding similarity computation with no self-referential definition. No self-citations are invoked as load-bearing uniqueness theorems or ansatzes, and the methodology is replicable against external benchmarks. The derivation chain therefore remains self-contained.

Axiom & Free-Parameter Ledger

2 free parameters · 2 axioms · 0 invented entities

The central claims rest on the assumption that job postings are a valid proxy for skill demand and that automated NLP pipelines extract meaningful categories without substantial measurement error. No new physical entities or forces are postulated.

free parameters (2)
  • Number of topics in LDA and KMeans
    Hyperparameter controlling granularity of skill clustering; value not stated in abstract but required for topic modeling results.
  • ARIMA model orders (p,d,q)
    Parameters fitted to time-series data for forecasting AI_Data and Soft_Meta skill growth.
axioms (2)
  • domain assumption Job postings language accurately reflects employer skill requirements without systematic bias from recruiter phrasing or platform norms.
    Invoked implicitly when treating extracted skills as direct evidence of workforce transformation.
  • domain assumption Sentence-transformer embeddings preserve semantic distinctions between augmentation and automation framing sufficiently for cosine similarity to yield a reliable Framing Index.
    Underlies the distinction between augmentative and substitutive discourse in the results.

pith-pipeline@v0.9.0 · 5562 in / 1651 out tokens · 31712 ms · 2026-05-10T18:23:40.311643+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

35 extracted references · 35 canonical work pages

  1. [1]

    The Impact of ChatGPT on HR Recruitment,

    Y. Zhang, “The Impact of ChatGPT on HR Recruitment,” J. Educ. Humanit. Soc. Sci., 2023, doi: 10.54097/ehss.v19i.10949

    work page doi:10.54097/ehss.v19i.10949 2023

  2. [2]

    Deep Learning Approaches for Big Data-Driven Metadata Extraction in Online Job Postings,

    P. Skondras, N. Zotos, D. Lagios, P. Zervas, K. C. Giotopoulos, and G. Tzimas, “Deep Learning Approaches for Big Data-Driven Metadata Extraction in Online Job Postings,” Inf., 2023, doi: 10.3390/info14110585

    work page doi:10.3390/info14110585 2023

  3. [3]

    Artificial Intelligence as a Service, Economic Growth, and Well- Being,

    C. A. Makridis and S. Mishra, “Artificial Intelligence as a Service, Economic Growth, and Well- Being,” J. Serv. Res., 2022, doi: 10.1177/10946705221120218

    work page doi:10.1177/10946705221120218 2022

  4. [4]

    Are online job postings informative to investors?,

    E. Gutiérrez, B. Lourie, A. Nekrasov, and T. Shevlin, “Are online job postings informative to investors?,” Manage. Sci., 2020, doi: 10.1287/mnsc.2019.3450

    work page doi:10.1287/mnsc.2019.3450 2020

  5. [5]

    Investigating the Demand for Blockchain Talents in the Recruitment Market: Evidence from Topic Modeling Analysis on Job Postings,

    C. GE, H. SHI, J. JIANG, and X. XU, “Investigating the Demand for Blockchain Talents in the Recruitment Market: Evidence from Topic Modeling Analysis on Job Postings,” Inf. Manag., 2021, doi: 10.1016/j.im.2021.103513

    work page doi:10.1016/j.im.2021.103513 2021

  6. [6]

    Identifying Multimedia Production Competencies and Skills of Instructional Design and Technology Professionals: An Analysis of Recent Job Postings,

    W. Sugar, B. Hoard, A. Brown, and L. Daniels, “Identifying Multimedia Production Competencies and Skills of Instructional Design and Technology Professionals: An Analysis of Recent Job Postings,” J. Educ. Technol. Syst., 2012, doi: 10.2190/et.40.3.b

    work page doi:10.2190/et.40.3.b 2012

  7. [7]

    Occupational models from 42 million unstructured job postings,

    N. Dixon et al., “Occupational models from 42 million unstructured job postings,” Patterns, 2023, doi: 10.1016/j.patter.2023.100757

    work page doi:10.1016/j.patter.2023.100757 2023

  8. [8]

    Ethical requirements in job advertisements: A deep learning approach,

    R. Liu, M. Li, S. Sin, and M. Tan, “Ethical requirements in job advertisements: A deep learning approach,” Eur. Manag. Rev., 2024, doi: 10.1111/emre.12569

    work page doi:10.1111/emre.12569 2024

  9. [9]

    Exercise professional education, qualifications, and certifications: a content analysis of job postings in the United States,

    R. Pojednic, D. P. O’Neill, M. G. Flanagan, A. Bartlett, B. L. G. Carter, and M. A. Kennedy, 22 “Exercise professional education, qualifications, and certifications: a content analysis of job postings in the United States,” Front. Sport. Act. Living, 2024, doi: 10.3389/fspor.2024.1338658

    work page doi:10.3389/fspor.2024.1338658 2024

  10. [10]

    Learning and development roles and competency domains in higher education: a content analysis of job announcements,

    F. Martin, Y. Chen, B. Oyarzun, and M. Lee, “Learning and development roles and competency domains in higher education: a content analysis of job announcements,” J. Comput. High. Educ., 2022, doi: 10.1007/s12528-021-09290-2

    work page doi:10.1007/s12528-021-09290-2 2022

  11. [11]

    Benefits, Limits, and Risks of GPT-4 as an AI Chatbot for Medicine.New England Journal of Medicine, 388(13):1233–1239, March 2023

    P. Lee, S. Bubeck, and J. Petro, “Benefits, Limits, and Risks of GPT-4 as an AI Chatbot for Medicine,” N. Engl. J. Med., 2023, doi: 10.1056/nejmsr2214184

    work page doi:10.1056/nejmsr2214184 2023

  12. [12]

    & Stephany, F

    E. Mäkelä and F. Stephany, “Complement or substitute? How AI increases the demand for human skills.” 2025. [Online]. Available: https://arxiv.org/abs/2412.19754

    work page arXiv 2025

  13. [13]

    Counteracting the Global Labor Shortage Risk Through the Human-AI Collaboration in Digital Recruiting,

    O. Linnyk and I. Teetz, “Counteracting the Global Labor Shortage Risk Through the Human-AI Collaboration in Digital Recruiting,” IEEE Technology and Society Magazine. 2023. doi: 10.1109/MTS.2023.3277108

    work page doi:10.1109/mts.2023.3277108 2023

  14. [14]

    Demand for AI Skills in Jobs: Evidence from Online Job Postings,

    M. Squicciarini and H. Nachtigall, “Demand for AI Skills in Jobs: Evidence from Online Job Postings,” OECD Sci. Technol. Ind. Work. Pap., 2021

    work page 2021

  15. [15]

    Enhancing Skills Demand Understanding through Job Ad Segmentation Using NLP and Clustering Techniques,

    M. Lukauskas, V. Šarkauskaitė, V. Pilinkienė, A. Stundžienė, A. Grybauskas, and J. Bruneckienė, “Enhancing Skills Demand Understanding through Job Ad Segmentation Using NLP and Clustering Techniques,” Appl. Sci., 2023, doi: 10.3390/app13106119

    work page doi:10.3390/app13106119 2023

  16. [16]

    Generative AI at Work

    E. Brynjolfsson, D. Li, and L. Raymond, “Generative-AI at Work.” 2024. [Online]. Available: https://arxiv.org/abs/2304.11771

    work page arXiv 2024

  17. [17]

    Science 381(6654): 187–92

    S. Noy and W. Zhang, “Experimental evidence on the productivity effects of generative artificial intelligence,” Science (80-. )., 2023, doi: 10.1126/science.adh2586

    work page doi:10.1126/science.adh2586 2023

  18. [18]

    When Combinations of Humans and

    M. Vaccaro, A. Almaatouq, and T. Malone, “When combinations of humans and AI are useful: A systematic review and meta-analysis,” Nat. Hum. Behav., vol. 8, no. 12, pp. 2293–2303, 2024, doi: 10.1038/s41562-024-02024-1

    work page doi:10.1038/s41562-024-02024-1 2024

  19. [19]

    Augmenting or Automating Labor? The Effect of AI Development on New Work, Employment, and Wages

    D. Marguerit, “Augmenting or Automating Labor? The Effect of AI Development on New Work, Employment, and Wages.” 2025. [Online]. Available: https://arxiv.org/abs/2503.19159

    work page arXiv 2025

  20. [20]

    AI exposure predicts unemployment risk: A new approach to technology-driven job loss,

    M. R. Frank, Y.-Y. Ahn, and E. Moro, “AI exposure predicts unemployment risk: A new approach to technology-driven job loss,” PNAS Nexus, vol. 4, no. 4, p. pgaf107, 2025, doi: 10.1093/pnasnexus/pgaf107

    work page doi:10.1093/pnasnexus/pgaf107 2025

  21. [21]

    Artificial intelligence and jobs: Evidence from online vacancies

    D. Acemoglu, D. Autor, J. Hazell, and P. Restrepo, “Artificial Intelligence and Jobs: Evidence from Online Vacancies,” J. Labor Econ., vol. 40, no. S1, pp. S293–S340, Apr. 2022, doi: 10.1086/718327

    work page doi:10.1086/718327 2022

  22. [22]

    Skills or degree? The rise of skill-based hiring for AI and green jobs,

    M. Bone, E. González Ehlinger, and F. Stephany, “Skills or degree? The rise of skill-based hiring for AI and green jobs,” Technol. Forecast. Soc. Change, vol. 214, p. 124042, 2025, doi: https://doi.org/10.1016/j.techfore.2025.124042

    work page doi:10.1016/j.techfore.2025.124042 2025

  23. [23]

    The paradoxes of generative-AI- enabled customer service: A guide for managers,

    C. Ferraro, V. Demsar, S. Sands, M. Restrepo, and C. Campbell, “The paradoxes of generative-AI- enabled customer service: A guide for managers,” Bus. Horiz., vol. 67, no. 5, pp. 549–559, 2024, doi: https://doi.org/10.1016/j.bushor.2024.04.013

    work page doi:10.1016/j.bushor.2024.04.013 2024

  24. [24]

    Algorithmic management in scientific research,

    M. Koehler and H. Sauermann, “Algorithmic management in scientific research,” Res. Policy, vol. 53, no. 4, p. 104985, 2024, doi: https://doi.org/10.1016/j.respol.2024.104985

    work page doi:10.1016/j.respol.2024.104985 2024

  25. [25]

    How generative-AI Is shaping the future of marketing,

    D. Grewal, C. B. Satornino, T. Davenport, and A. Guha, “How generative-AI Is shaping the future of marketing,” J. Acad. Mark. Sci., vol. 53, no. 3, pp. 702–722, 2025, doi: 10.1007/s11747-024- 01064-3

    work page doi:10.1007/s11747-024- 2025

  26. [26]

    Generative artificial intelligence usage by researchers at work: Effects of gender, career stage, type of workplace, and perceived barriers,

    P. Dorta-González, A. J. López-Puig, M. I. Dorta-González, and S. M. González-Betancor, “Generative artificial intelligence usage by researchers at work: Effects of gender, career stage, type of workplace, and perceived barriers,” Telemat. Informatics, vol. 94, p. 102187, 2024, doi: https://doi.org/10.1016/j.tele.2024.102187

    work page doi:10.1016/j.tele.2024.102187 2024

  27. [27]

    Generative artificial intelligence, human creativity, and art.,

    E. Zhou and D. Lee, “Generative artificial intelligence, human creativity, and art.,” PNAS nexus, vol. 3, no. 3, p. pgae052, Mar. 2024, doi: 10.1093/pnasnexus/pgae052

    work page doi:10.1093/pnasnexus/pgae052 2024

  28. [28]

    The ethics of ChatGPT in medicine and healthcare: a systematic review on Large Language Models (LLMs),

    J. Haltaufderheide and R. Ranisch, “The ethics of ChatGPT in medicine and healthcare: a systematic review on Large Language Models (LLMs),” npj Digit. Med., vol. 7, no. 1, p. 183, 23 2024, doi: 10.1038/s41746-024-01157-x

    work page doi:10.1038/s41746-024-01157-x 2024

  29. [29]

    A systematic review and meta -analysis of diagnostic performance comparison between generative AI and physicians,

    H. Takita et al., “A systematic review and meta-analysis of diagnostic performance comparison between generative-AI and physicians,” npj Digit. Med., vol. 8, no. 1, p. 175, 2025, doi: 10.1038/s41746-025-01543-z

    work page doi:10.1038/s41746-025-01543-z 2025

  30. [30]

    The Double-Edged Roles of Generative-AI in the Creative Process: Experiments on Design Work,

    J. (Jove) Hou, L. Wang, G. Wang, H. J. Wang, and S. Yang, “The Double-Edged Roles of Generative-AI in the Creative Process: Experiments on Design Work,” Inf. Syst. Res., Oct. 2025, doi: 10.1287/isre.2024.0937

    work page doi:10.1287/isre.2024.0937 2025

  31. [31]

    Generative-AI and its Transformative Value for Digital Platforms,

    M. Wessel, M. Adam, A. Benlian, A. Majchrzak, and F. Thies, “Generative-AI and its Transformative Value for Digital Platforms,” J. Manag. Inf. Syst., vol. 42, no. 2, pp. 346–369, Apr. 2025, doi: 10.1080/07421222.2025.2487315

    work page doi:10.1080/07421222.2025.2487315 2025

  32. [32]

    Scientific Reports15(1), 13755 (2025)

    M. Raza, Z. Jahangir, M. B. Riaz, M. J. Saeed, and M. A. Sattar, “Industrial applications of large language models,” Sci. Rep., vol. 15, no. 1, p. 13755, 2025, doi: 10.1038/s41598-025-98483-1

    work page doi:10.1038/s41598-025-98483-1 2025

  33. [33]

    Artificial intelligence and the wellbeing of workers,

    O. Giuntella, J. Konig, and L. Stella, “Artificial intelligence and the wellbeing of workers,” Sci. Rep., vol. 15, no. 1, p. 20087, 2025, doi: 10.1038/s41598-025-98241-3

    work page doi:10.1038/s41598-025-98241-3 2025

  34. [34]

    Enhancing Work Productivity through Generative Artificial Intelligence: A Comprehensive Literature Review,

    H. Al Naqbi, Z. Bahroun, and V. Ahmed, “Enhancing Work Productivity through Generative Artificial Intelligence: A Comprehensive Literature Review,” Sustainability, vol. 16, no. 3, 2024, doi: 10.3390/su16031166

    work page doi:10.3390/su16031166 2024

  35. [35]

    How artificial intelligence affects the labour force employment structure from the perspective of industrial structure optimisation,

    X. Wang, M. Chen, and N. Chen, “How artificial intelligence affects the labour force employment structure from the perspective of industrial structure optimisation,” Heliyon, 2024, doi: 10.1016/j.heliyon.2024.e26686

    work page doi:10.1016/j.heliyon.2024.e26686 2024