pith. sign in

arxiv: 2512.03671 · v2 · pith:R2PVYABWnew · submitted 2025-12-03 · 💻 cs.CL

Generative AI Practices, Literacy, and Divides: An Empirical Analysis in the Italian Context

Beatrice Savoldi , Giuseppe Attanasio , Olga Gorodetskaya , Marta Marchiori Manerba , Elisa Bassignana , Silvia Casola , Matteo Negri , Tommaso Caselli
show 5 more authors
Luisa Bentivogli Alan Ramponi Arianna Muti Nicoletta Balbo Debora Nozza
This is my paper

Pith reviewed 2026-05-21 18:20 UTC · model grok-4.3

classification 💻 cs.CL
keywords generative AIdigital literacytechnology adoptiondigital dividessurvey analysisAI trainingusage patternsItaly
0
0 comments X

The pith

AI training is the main factor that turns generative AI use into content creation and learning rather than passive recreation.

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

This paper draws on a survey of 1,906 Italian-speaking adults to map who adopts generative AI chatbots and how they actually use them. Adoption is lower among less-educated, older, and less tech-familiar people, and 40 percent cite lack of competence as a barrier. Among those who do use the tools, prior AI training strongly predicts engagement in productive activities such as creating content, supporting learning, and boosting creativity. Passive or recreational uses, by contrast, show up at similar rates regardless of training or competence. The results indicate that simple accessibility does not automatically spread the benefits of these tools evenly.

Core claim

Drawing on original survey data from 1,906 Italian-speaking adults, the analysis shows that generative AI supports diversified personal and professional activities and is replacing traditional information-seeking tools. Less-educated and older individuals and those with lower technology familiarity adopt it less often, with 40 percent citing competence barriers. Among users, AI training emerges as the primary predictor of purposeful, capital-enhancing engagement such as content creation, learning, and creativity enhancement, while more passive recreational uses remain insensitive to competence levels. Gender operates as a persistent cross-cutting divide in both adoption and usage frequency.

What carries the argument

AI training as the predictor that separates purposeful capital-enhancing GenAI engagement from passive recreational uses among adopters.

If this is right

  • Programs that supply AI training can raise the share of productive GenAI activities beyond what access alone achieves.
  • Digital divides in the GenAI era concern the quality and purpose of use, not only whether people start using the tools.
  • Gender differences continue to shape both the choice to adopt and how often people interact with generative AI.
  • Policies aimed at literacy can address emerging gaps in who gains from the technology.

Where Pith is reading between the lines

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

  • Comparable competence gaps in productive use may appear in other countries that share similar patterns of education and technology access.
  • Targeted training experiments could test whether specific literacy interventions increase rates of creative and educational applications over time.
  • Persistent differences in usage purpose may widen longer-term skill and economic gaps if left unaddressed.

Load-bearing premise

Self-reported survey answers accurately reflect actual GenAI usage and the Italian adult sample is representative enough to support general claims about divides.

What would settle it

Direct logs or observation of GenAI interactions showing whether users who received training actually produce more content and learning activities than untrained users of similar age and education.

read the original abstract

The rise of generative AI (GenAI) chatbots accessible via conversational interfaces is transforming digital interactions and holds economic promise. However, these tools might deepen existing inequalities -- not only through uneven, socially stratified adoption, but through differentials in their purposeful, critical use. Drawing on original survey data from 1,906 Italian-speaking adults, we provide a comprehensive analysis of GenAI adoption, literacy, and usage patterns. Our findings show that GenAI is supporting diversified personal and professional activities and replacing traditional information-seeking tools. Yet less-educated and older individuals, and those with lower technology familiarity, are less likely to adopt it; 40% cite competence barriers as a key obstacle. Among users, AI training emerges as the primary predictor of purposeful, capital-enhancing engagement -- content creation, learning, and creativity enhancement -- while more passive, recreational uses (e.g., companionship, information seeking) remain insensitive to competence levels. We thus highlight digital literacy as a lever for how people leverage GenAI, not just whether they use it. Finally, gender operates as a persistent cross-cutting divide, shaping both adoption and usage frequency. These findings challenge the assumption that high accessibility translates into broadly shared gains. Rather, they offer a granular, multi-level account of emerging disparities in the GenAI era -- with implications for how this technology may ultimately drive outcomes and benefit divides.

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

2 major / 2 minor

Summary. The paper presents findings from an original survey of 1,906 Italian-speaking adults on generative AI adoption, literacy, and usage patterns. It reports stratified adoption (lower among less-educated, older, and lower-tech-familiarity respondents), with 40% citing competence barriers; among users, AI training is identified as the primary predictor of purposeful, capital-enhancing activities (content creation, learning, creativity) while recreational uses remain insensitive to competence; gender is a cross-cutting divide. The work concludes that digital literacy shapes how GenAI is leveraged, not merely whether it is adopted, challenging assumptions of broad accessibility benefits.

Significance. If the methodological details and robustness checks are supplied, the study offers a timely, granular empirical account of emerging GenAI divides in a national context. The sample size of 1,906 supports descriptive and correlational claims about usage patterns and the role of training; explicit credit is due for the multi-level framing that separates adoption from purposeful use and for highlighting training as a potential policy lever.

major comments (2)
  1. [Methods] Methods section (sampling and data collection): The manuscript provides no information on recruitment method, response rate, or weighting to ensure representativeness of the 1,906 Italian-speaking adults. This is load-bearing for all claims about population-level divides by education, age, and gender, as convenience or online-panel sampling could systematically under-represent low-literacy or older non-users and compress variance in the very disparities documented.
  2. [Results] Results section (predictor analysis): The central claim that 'AI training emerges as the primary predictor of purposeful, capital-enhancing engagement' rests on self-reported measures of both training and usage types without reported validation, behavioral logs, or controls for common-method bias. Because training and outcome categories are collected via the same instrument, recall and social-desirability biases could induce spurious associations; the abstract and results give no indication of robustness checks that would separate these effects.
minor comments (2)
  1. [Abstract] Abstract: The phrase 'primary predictor' should be qualified with the specific statistical approach (e.g., regression coefficients, effect sizes, or descriptive cross-tabs) used to establish primacy over other variables.
  2. [Results] The manuscript would benefit from a table summarizing response distributions by key demographics to allow readers to assess the observed divides directly.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive and detailed feedback on our manuscript. We address each major comment below, indicating where revisions will be made to strengthen the paper.

read point-by-point responses

  1. Referee: [Methods] Methods section (sampling and data collection): The manuscript provides no information on recruitment method, response rate, or weighting to ensure representativeness of the 1,906 Italian-speaking adults. This is load-bearing for all claims about population-level divides by education, age, and gender, as convenience or online-panel sampling could systematically under-represent low-literacy or older non-users and compress variance in the very disparities documented.

    Authors: We agree that the current Methods section is insufficiently detailed on these points, which limits the interpretability of our population-level claims. In the revised manuscript we will expand this section to describe the recruitment process (conducted via a commercial online panel with targeted invitations to Italian-speaking adults), report the response rate and any eligibility screening, and detail the post-stratification weighting applied to match national benchmarks on age, gender, education, and geographic region. These additions will directly address concerns about potential under-representation of lower-literacy or older non-users. revision: yes

  2. Referee: [Results] Results section (predictor analysis): The central claim that 'AI training emerges as the primary predictor of purposeful, capital-enhancing engagement' rests on self-reported measures of both training and usage types without reported validation, behavioral logs, or controls for common-method bias. Because training and outcome categories are collected via the same instrument, recall and social-desirability biases could induce spurious associations; the abstract and results give no indication of robustness checks that would separate these effects.

    Authors: We acknowledge the risk of common-method bias inherent to single-instrument self-report data. In the revision we will add an explicit discussion of this limitation, include additional robustness checks (e.g., models with demographic and technology-familiarity controls, and sensitivity analyses separating training from usage items), and clarify that the primary-predictor result emerges from multivariate regressions rather than bivariate associations. While we cannot retroactively supply behavioral logs or external validation, we will strengthen the presentation of these correlational findings and their policy implications. revision: partial

Circularity Check

0 steps flagged

No circularity: purely descriptive survey analysis with no derivations or self-referential reductions

full rationale

The paper reports results from an original survey of 1,906 Italian-speaking adults on GenAI adoption, literacy, barriers, and usage patterns. All claims, including the identification of AI training as the primary predictor of purposeful engagement, rest on direct tabulation and correlation of self-reported responses. No equations, fitted models presented as out-of-sample predictions, uniqueness theorems, or ansatzes appear in the provided text. The analysis does not reduce any result to its own inputs by construction, nor does it rely on load-bearing self-citations that substitute for independent evidence. This is a standard empirical survey study whose internal logic is self-contained against the collected data.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The central claims rest on standard survey assumptions rather than new theoretical constructs. No free parameters or invented entities are introduced.

axioms (2)
  • domain assumption Survey responses on self-reported usage, barriers, and training levels validly reflect actual behaviors and competence.
    Invoked when interpreting competence barriers and training as predictors of purposeful use.
  • domain assumption The 1,906-respondent sample is representative of Italian-speaking adults for purposes of identifying population-level divides.
    Required to generalize findings beyond the surveyed group.

pith-pipeline@v0.9.0 · 5829 in / 1094 out tokens · 72301 ms · 2026-05-21T18:20:12.128287+00:00 · methodology

discussion (0)

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

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Forward citations

Cited by 6 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. From Chatbots to Confidants: A Cross-Cultural Study of LLM Adoption for Emotional Support

    cs.CL 2026-04 unverdicted novelty 7.0

    A cross-cultural survey finds LLM emotional support adoption ranges from 20% to 59% by country, with positive perceptions strongest among higher-SES, religious, married adults aged 25-44 and in English-speaking nations.

  2. From Chatbots to Confidants: A Cross-Cultural Study of LLM Adoption for Emotional Support

    cs.CL 2026-04 unverdicted novelty 7.0

    Cross-cultural survey of 4,641 participants shows LLM emotional support adoption varies widely by country and demographics, with socioeconomic status as strongest predictor of trust and use, and English-speaking natio...

  3. From Exposure to Adoption: Generative AI in European Workplaces

    econ.GN 2026-04 unverdicted novelty 5.0

    Generative AI adoption in Europe ranges from under 3% to 25%, is steeper for skilled workers in abstract-task jobs and in digitally advanced countries with training, shows a gender gap in exposed roles, and has produc...

  4. A Techno-Economic Framework for Cost Modeling and Revenue Opportunities in Open and Programmable AI-RAN

    cs.NI 2026-03 unverdicted novelty 5.0

    GPU-accelerated RAN can achieve up to 8x ROI by leasing surplus compute capacity to AI tenants, offsetting added capital and operational costs.

  5. A Techno-Economic Framework for Cost Modeling and Revenue Opportunities in Open and Programmable AI-RAN

    cs.NI 2026-03 unverdicted novelty 5.0

    Techno-economic framework shows that GPU AI-RAN deployments can offset extra costs via AI revenue for up to 8x ROI across scenarios with varying token depreciation, demand, and GPU densities.

  6. Puzzled By ChatGPT? No more! A Jigsaw Puzzle to Promote AI Literacy and Awareness

    cs.CL 2026-05 unverdicted novelty 3.0

    A jigsaw puzzle with comic-based infographics is presented as an interactive tool to promote public understanding of generative AI systems like ChatGPT.

Reference graph

Works this paper leans on

63 extracted references · 63 canonical work pages · cited by 4 Pith papers

  1. [1]

    SN Computer Science 5(8):1154

    Abbes F, Bennani S, Maalel A (2024) Generative AI and gamification for personalized learning: Literature review and future challenges . SN Computer Science 5(8):1154

    work page 2024

  2. [2]

    Scientific Reports 14

    Alipour S, Galeazzi A, Sangiorgio E, et al (2023) Cross-platform social dynamics: an analysis of chatgpt and covid-19 vaccine conversations. Scientific Reports 14. ://api.semanticscholar.org/CorpusID:264172259

    work page 2023

  3. [3]

    Humanities and Social Sciences Communications 12(1):1--7

    Andersson M (2025) Companionship in code: AI’s role in the future of human connection . Humanities and Social Sciences Communications 12(1):1--7

    work page 2025

  4. [4]

    In: Che W, Nabende J, Shutova E, et al (eds) Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)

    Bassignana E, Curry AC, Hovy D (2025) The AI gap: How socioeconomic status affects language technology interactions. In: Che W, Nabende J, Shutova E, et al (eds) Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). Association for Computational Linguistics, Vienna, Austria, pp 18647--18664, doi:1...

    work page doi:10.18653/v1/2025.acl-long.914 2025

  5. [5]

    Beccalli E, Pais I, Rosina A, et al (eds) (2024) Intelligenza artificiale: rischi e opportunit \`a . No. 11 in Quaderni Istituto Toniolo - Rapporto Giovani, Vita e Pensiero / Istituto Toniolo, Milan

    work page 2024

  6. [6]

    AI and Ethics 4(4):901--915

    Bentley SV, Naughtin CK, McGrath MJ, et al (2024) The digital divide in action: How experiences of digital technology shape future relationships with Artificial Intelligence . AI and Ethics 4(4):901--915

    work page 2024

  7. [7]

    arXiv preprint arXiv:251007557

    Bhandarkar A, Mishra G, Juchani K, et al (2025) Investigating thematic patterns and user preferences in llm interactions using bertopic. arXiv preprint arXiv:251007557

    work page 2025

  8. [8]

    ://arxiv.org/abs/2402.02136, https://arxiv.org/abs/2402.02136 arXiv:2402.02136

    Bodonhelyi A, Bozkir E, Yang S, et al (2024) User intent recognition and satisfaction with large language models: A user study with chatgpt. ://arxiv.org/abs/2402.02136, https://arxiv.org/abs/2402.02136 arXiv:2402.02136

    work page arXiv 2024

  9. [9]

    CoRR abs/2507.05046

    Bouyzourn K, Birch A (2025) What shapes user trust in chatgpt? A mixed-methods study of user attributes, trust dimensions, task context, and societal perceptions among university students. CoRR abs/2507.05046

    work page arXiv 2025

  10. [10]

    The Quarterly Journal of Economics 140(2):889--942

    Brynjolfsson E, Li D, Raymond L (2025) Generative AI at work . The Quarterly Journal of Economics 140(2):889--942

    work page 2025

  11. [11]

    PNAS nexus 3(6):pgae191

    Capraro V, Lentsch A, Acemoglu D, et al (2024) The impact of generative artificial intelligence on socioeconomic inequalities and policy making. PNAS nexus 3(6):pgae191

    work page 2024

  12. [12]

    Carvajal D, Franco C, Isaksson S (2024) Will Artificial Intelligence get in the way of achieving gender equality? NHH Dept of Economics Discussion Paper

    work page 2024

  13. [13]

    Chatterji A, Cunningham T, Deming DJ, et al (2025) How people use ChatGPT . Tech. rep., National Bureau of Economic Research

    work page 2025

  14. [14]

    AI Magazine 45(3):354--368

    Chen C, Shu K (2024) Combating misinformation in the age of LLMs: Opportunities and challenges . AI Magazine 45(3):354--368

    work page 2024

  15. [15]

    ://www.commonsensemedia.org/sites/default/files/research/report/talk-trust-and-trade-offs_2025_web.pdf

    Common Sense Media (2025) Talk, trust, and trade-offs: How and why teens use ai companion. ://www.commonsensemedia.org/sites/default/files/research/report/talk-trust-and-trade-offs_2025_web.pdf

    work page 2025

  16. [16]

    ://arxiv.org/abs/2404.11988, https://arxiv.org/abs/2404.11988 arXiv:2404.11988

    Daepp MIG, Counts S (2025) The emerging generative artificial intelligence divide in the united states. ://arxiv.org/abs/2404.11988, https://arxiv.org/abs/2404.11988 arXiv:2404.11988

    work page arXiv 2025

  17. [17]

    ://www.primaonline.it/wp-content/uploads/2025/05/eurispes_-sintesi-37-rapporto-italia_compressed.pdf

    Eurispes (2025) 37° rapporto italia. ://www.primaonline.it/wp-content/uploads/2025/05/eurispes_-sintesi-37-rapporto-italia_compressed.pdf

    work page 2025

  18. [18]

    https://explodingtopics.com/blog/most-visited-websites, accessed: October 15, 2025

    Exploding Topics (2025) Most Visited Websites In The World (August 2025) . https://explodingtopics.com/blog/most-visited-websites, accessed: October 15, 2025

    work page 2025

  19. [19]

    In: Ageing and digital technology: Designing and evaluating emerging technologies for older adults

    Francis J, Ball C, Kadylak T, et al (2019) Aging in the digital age: Conceptualizing technology adoption and digital inequalities . In: Ageing and digital technology: Designing and evaluating emerging technologies for older adults. Springer, p 35--49

    work page 2019

  20. [20]

    University of Melbourne, KPMG

    Gillespie N, Lockey S, Ward T, et al (2025) Trust, attitudes and use of artificial intelligence. University of Melbourne, KPMG

    work page 2025

  21. [21]

    ://www.theguardian.com/technology/2025/oct/16/google-ai-overviews-italian-news-publishers-demand-investigation, accessed: 2025-10-19

    Giuffrida A (2025) Google ai overviews: Italian news publishers demand investigation. ://www.theguardian.com/technology/2025/oct/16/google-ai-overviews-italian-news-publishers-demand-investigation, accessed: 2025-10-19

    work page 2025

  22. [22]

    arXiv preprint arXiv:220305794

    Grootendorst M (2022) Bertopic: Neural topic modeling with a class-based tf-idf procedure. arXiv preprint arXiv:220305794

    work page 2022

  23. [23]

    arXiv preprint arXiv:250304761

    Handa K, Tamkin A, McCain M, et al (2025) Which economic tasks are performed with AI? Evidence from millions of Claude conversations . arXiv preprint arXiv:250304761

    work page 2025

  24. [24]

    Available at SSRN

    Hartley J, Jolevski F, Melo V, et al (2024) The labor market effects of generative artificial intelligence. Available at SSRN

    work page 2024

  25. [25]

    Nature 633(8028):147--154

    Hofmann V, Kalluri PR, Jurafsky D, et al (2024) Ai generates covertly racist decisions about people based on their dialect. Nature 633(8028):147--154

    work page 2024

  26. [26]

    In: Ku LW, Martins A, Srikumar V (eds) Findings of the Association for Computational Linguistics: ACL 2024

    Holtermann C, R \"o ttger P, Dill T, et al (2024) Evaluating the elementary multilingual capabilities of large language models with M ulti Q . In: Ku LW, Martins A, Srikumar V (eds) Findings of the Association for Computational Linguistics: ACL 2024. Association for Computational Linguistics, Bangkok, Thailand, pp 4476--4494, doi:10.18653/v1/2024.findings...

    work page doi:10.18653/v1/2024.findings-acl.265 2024

  27. [27]

    ://www.technologyreview.com/2025/01/06/1108679/ai-generative-search-internet-breakthroughs/

    Honan M (2025) AI means the end of internet search as we've known it . ://www.technologyreview.com/2025/01/06/1108679/ai-generative-search-internet-breakthroughs/

    work page 2025

  28. [28]

    ://resources.ipsos.com/rs/297-CXJ-795/images/Ipsos-AI-Monitor-2025.pdf

    Ipsos (2025) The ipsos ai monitor 2025: A 30-country ipsos global advisor survey. ://resources.ipsos.com/rs/297-CXJ-795/images/Ipsos-AI-Monitor-2025.pdf

    work page 2025

  29. [29]

    Digital Skills and Jobs Platform, ://digital-skills-jobs.europa.eu/en/latest/briefs/italy-snapshot-digital-skills, accessed: 2025-10-24

    Kralj L (2024) Italy: A snapshot of digital skills. Digital Skills and Jobs Platform, ://digital-skills-jobs.europa.eu/en/latest/briefs/italy-snapshot-digital-skills, accessed: 2025-10-24

    work page 2024

  30. [30]

    arXiv preprint arXiv:230409339

    Kreitmeir DH, Raschky PA (2023) The unintended consequences of censoring digital technology--evidence from italy's chatgpt ban. arXiv preprint arXiv:230409339

    work page 2023

  31. [31]

    Journal of experimental political science 9(1):104--117

    Kreps S, McCain RM, Brundage M (2022) All the news that’s fit to fabricate: AI-generated text as a tool of media misinformation . Journal of experimental political science 9(1):104--117

    work page 2022

  32. [32]

    In: Ku LW, Martins A, Srikumar V (eds) Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)

    Li J, Chen J, Ren R, et al (2024) The Dawn After the Dark: An Empirical Study on Factuality Hallucination in Large Language Models . In: Ku LW, Martins A, Srikumar V (eds) Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). Association for Computational Linguistics, Bangkok, Thailand, pp 10879--...

    work page doi:10.18653/v1/2024.acl-long.586 2024

  33. [33]

    Mitchell M, Attanasio G, Baldini I, et al (2025) SHADES: Towards a multilingual assessment of stereotypes in large language models . In: Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers), pp 11995--12041

    work page 2025

  34. [34]

    Education Sciences 14(12):1363

    M gelvang A, Bjelland C, Grassini S, et al (2024) Gender differences in the use of generative artificial intelligence chatbots in higher education: Characteristics and consequences. Education Sciences 14(12):1363

    work page 2024

  35. [35]

    In: Proceedings of the International AAAI Conference on Web and Social Media, pp 1262--1294

    Murayama T, Miyazaki K, Matsubara Y, et al (2025) Linguistic Landscape of Generative AI Perception: A Global Twitter Analysis Across 14 Languages . In: Proceedings of the International AAAI Conference on Web and Social Media, pp 1262--1294

    work page 2025

  36. [36]

    Science 381(6654):187--192

    Noy S, Zhang W (2023) Experimental evidence on the productivity effects of generative artificial intelligence . Science 381(6654):187--192

    work page 2023

  37. [37]

    https://chat.openai.com, accessed: October 15, 2025

    OpenAI (2023) ChatGPT : Large language model. https://chat.openai.com, accessed: October 15, 2025

    work page 2023

  38. [38]

    Harvard Business School Boston, MA, USA

    Otis NG, Delecourt S, Cranney K, et al (2024) Global evidence on gender gaps and generative AI. Harvard Business School Boston, MA, USA

    work page 2024

  39. [39]

    In: Bouamor H, Pino J, Bali K (eds) Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing

    Ouyang S, Wang S, Liu Y, et al (2023) The shifted and the overlooked: A task-oriented investigation of user- GPT interactions. In: Bouamor H, Pino J, Bali K (eds) Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing. Association for Computational Linguistics, Singapore, pp 2375--2393, doi:10.18653/v1/2023.emnlp-main.146, ...

    work page doi:10.18653/v1/2023.emnlp-main.146 2023

  40. [40]

    Quality & quantity 41(5):673--690

    O’brien RM (2007) A caution regarding rules of thumb for variance inflation factors . Quality & quantity 41(5):673--690

    work page 2007

  41. [41]

    Pew Research Center (2025 a ) 34\

    work page 2025

  42. [42]

    Pew Research Center (2025 b ) Do people click on links in google ai summaries? ://www.pewresearch.org/short-reads/2025/07/22/google-users-are-less-likely-to-click-on-links-when-an-ai-summary-appears-in-the-results/

    work page 2025

  43. [43]

    public and ai experts view artificial intelligence

    Pew Research Center (2025 c ) How the u.s. public and ai experts view artificial intelligence. ://www.pewresearch.org/internet/2025/04/03/how-the-us-public-and-ai-experts-view-artificial-intelligence/

    work page 2025

  44. [44]

    ://www.qualtrics.com, software

    Qualtrics (2025) Qualtrics xm. ://www.qualtrics.com, software

    work page 2025

  45. [45]

    IEEE Transactions on Artificial Intelligence

    Scantamburlo T, Cort \'e s A, Foffano F, et al (2024) Artificial intelligence across europe: A study on awareness, attitude and trust . IEEE Transactions on Artificial Intelligence

    work page 2024

  46. [46]

    Telecommunications Policy 48(10):102870

    Serrecchia M (2024) Analysis of internet development and internal digital divide by using the “.it” domain names as an indicator: Evidence from italy. Telecommunications Policy 48(10):102870. doi:https://doi.org/10.1016/j.telpol.2024.102870, ://www.sciencedirect.com/science/article/pii/S0308596124001678

    work page doi:10.1016/j.telpol.2024.102870 2024

  47. [47]

    ACM Trans Web 19(3)

    Shah C, White R, Andersen R, et al (2025) Using large language models to generate, validate, and apply user intent taxonomies. ACM Trans Web 19(3). doi:10.1145/3732294, ://doi.org/10.1145/3732294

    work page doi:10.1145/3732294 2025

  48. [48]

    First Monday 29(1)

    Skjuve M, Brandtzaeg PB, F lstad A (2024) Why do people use ChatGPT? Exploring user motivations for generative conversational AI . First Monday 29(1)

    work page 2024

  49. [49]

    arXiv preprint arXiv:241217486

    Sublime J, Renna I (2024) Is chatgpt massively used by students nowadays? a survey on the use of large language models such as chatgpt in educational settings. arXiv preprint arXiv:241217486

    work page 2024

  50. [50]

    doi:https://doi.org/10.1016/j.telpol.2025.102997, ://www.sciencedirect.com/science/article/pii/S0308596125000941

    Suárez D, García-Mariñoso B (2025) On the verge of a digital divide in the use of generative AI? Telecommunications Policy 49(7):102997. doi:https://doi.org/10.1016/j.telpol.2025.102997, ://www.sciencedirect.com/science/article/pii/S0308596125000941

    work page doi:10.1016/j.telpol.2025.102997 2025

  51. [51]

    ://arxiv.org/abs/2507.07935, https://arxiv.org/abs/2507.07935 arXiv:2507.07935

    Tomlinson K, Jaffe S, Wang W, et al (2025) Working with AI: Measuring the Applicability of Generative AI to Occupations . ://arxiv.org/abs/2507.07935, https://arxiv.org/abs/2507.07935 arXiv:2507.07935

    work page arXiv 2025

  52. [52]

    In: Proceedings of the 47th International ACM SIGIR Conference on Research and Development in Information Retrieval, pp 2703--2707

    Trippas JR, Al Lawati SFD, Mackenzie J, et al (2024) What do users really ask large language models? An initial log analysis of Google Bard interactions in the wild . In: Proceedings of the 47th International ACM SIGIR Conference on Research and Development in Information Retrieval, pp 2703--2707

    work page 2024

  53. [53]

    Journal of Marketing p 00222429251314491

    Tully SM, Longoni C, Appel G (2025) Lower artificial intelligence literacy predicts greater ai receptivity. Journal of Marketing p 00222429251314491

    work page 2025

  54. [54]

    https://www.unesco.org/en/articles/ai-literacy-and-new-digital-divide-global-call-action, accessed: October 15, 2025

    UNESCO (2024) AI literacy and the new digital divide: A global call for action. https://www.unesco.org/en/articles/ai-literacy-and-new-digital-divide-global-call-action, accessed: October 15, 2025

    work page 2024

  55. [55]

    The Sage handbook of survey methods

    Vehovar V, Toepoel V, Steinmetz S (2016) Non-probability sampling , vol 1. The Sage handbook of survey methods

    work page 2016

  56. [56]

    Wang C, Boerman SC, Kroon AC, et al (2025) The artificial intelligence divide: Who is the most vulnerable? New Media & Society 27(7):3867--3889

    work page 2025

  57. [57]

    In: Bouamor H, Pino J, Bali K (eds) Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing

    Zhang X, Li S, Hauer B, et al (2023) Don ' t trust C hat GPT when your question is not in E nglish: A study of multilingual abilities and types of LLM s. In: Bouamor H, Pino J, Bali K (eds) Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing. Association for Computational Linguistics, Singapore, pp 7915--7927, doi:10.186...

    work page doi:10.18653/v1/2023.emnlp-main.491 2023

  58. [58]

    Computational Linguistics pp 1--46

    Zhang Y, Li Y, Cui L, et al (2025) Siren’s song in the ai ocean: A survey on hallucination in large language models. Computational Linguistics pp 1--46

    work page 2025

  59. [59]

    ://arxiv.org/abs/2510.15951, https://arxiv.org/abs/2510.15951 arXiv:2510.15951

    Zhou K, Gligorić K, Cheng M, et al (2025) Attention to Non-Adopters . ://arxiv.org/abs/2510.15951, https://arxiv.org/abs/2510.15951 arXiv:2510.15951

    work page arXiv 2025

  60. [60]

    sn-basic.bst

    FUNCTION identify.basic.version "sn-basic.bst" " [2024/07/19 v1.1 bibliography style]" * top ENTRY address archive author booktitle chapter doi edition editor eid eprint howpublished institution journal key keywords month note number organization pages publisher school series title type url volume year archivePrefix primaryClass adsurl adsnote version lab...

    work page 2024

  61. [61]

    write newline

    " write newline "" before.all 'output.state := FUNCTION add.period duplicate empty 'skip "." * add.blank if FUNCTION if.digit duplicate "0" = swap duplicate "1" = swap duplicate "2" = swap duplicate "3" = swap duplicate "4" = swap duplicate "5" = swap duplicate "6" = swap duplicate "7" = swap duplicate "8" = swap "9" = or or or or or or or or or FUNCTION ...

    work page

  62. [62]

    sn-nature.bst

    FUNCTION identify.nature.version "sn-nature.bst" " [2024/07/19 v1.1 bibliography style]" * top ENTRY address archive author booktitle chapter edition editor eprint howpublished institution journal key keywords month note number organization pages publisher school series title type url doi volume year archivePrefix primaryClass eid adsurl adsnote version l...

    work page 2024

  63. [63]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in capitalize " " * FUNCT...

    work page