{"id":"303e6f7b-562c-4f35-9380-765b019ecbd7","arxiv_id":"2606.17887","paper_version":1,"verdict":"CONDITIONAL","confidence":"LOW","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"GenAI adoption in HR depends on sociotechnical fit with employee positionalities and trust mechanisms such as source verification and colleague consultation.","lead":"Researchers analyzed logs, a survey of 25 employees, and 10 interviews during one tech company's switch from a legacy HR search tool to a GenAI version. Adoption varied by employees' roles, languages, and tenure, with trust built mainly through source checks and colleague input.","discovery_kind":"new_application","skeptic_critique":{"model":"grok-4.3","headline":"Single-organization case (n=25 survey + 10 interviews) supplies the only evidence for positionalities (role/language/tenure) as adoption drivers","rationale":"The reader’s weakest_assumption correctly isolates the single point where the argument’s scope exceeds its empirical base. No deeper internal inconsistency (e.g., contradictory coding or missing variable) is visible from the supplied material; the limitation is simply the narrow sampling frame.","tokens_in":1784,"tokens_out":283,"duration_ms":13521,"concrete_test":"Re-analyze the search-log data for adoption metrics (query volume, session length, follow-up clicks) stratified by the same positionalities (role, language, tenure) reported in the interviews; if the log-derived adoption rates show no statistically detectable differences across those strata, the positionalities claim lacks independent support.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim requires that observed patterns of fit between system assumptions and employee positionalities, plus trust-calibration behaviors, are sufficiently stable to ground design recommendations for high-stakes HR GenAI. The study supplies no quantitative triangulation from the search-log data, no cross-site replication, and no explicit test of whether the same positionalities predict adoption outside this firm; therefore the leap from this case to “design considerations for inclusive deployment” rests on untested generalizability.","agreement_with_reader":"agree"},"referee_report":{"model":"grok-4.3","summary":"The paper reports a mixed-methods case study of GenAI adoption during a live transition from a legacy HR search system to a GenAI-supported system at one multinational tech company. Drawing on search logs, a survey (n=25), and 10 semi-structured interviews, it claims that adoption is shaped by the fit between the system's design assumptions and employees' positionalities (role, spoken language, tenure), and that trust is constructed via source-checking, cross-system comparison, and consultation with colleagues or HR. The stated contributions are empirical evidence on sociotechnical adoption factors (situational fit, search literacy, trust calibration, knowledge infrastructure) plus design considerations for inclusive GenAI deployment in high-stakes HR settings.","tokens_in":1891,"tokens_out":494,"duration_ms":18771,"significance":"If the patterns hold, the work supplies timely, real-world evidence on how positionalities and trust-calibration practices affect workplace GenAI uptake, which could usefully inform HCI and CSCW research on sociotechnical systems. The live-transition setting and mixed-methods design are strengths. However, the single-site, small-sample evidence base (explicitly noted as n=25 survey + 10 interviews) substantially constrains the strength of any broader claims about adoption dynamics or design recommendations.","major_comments":[{"comment":"Methods and Findings sections: The abstract and introduction state that search log data were analyzed alongside the survey and interviews, yet no quantitative results, metrics, or triangulation from the logs are reported to corroborate the positionalities (role/language/tenure) claims; the central findings therefore rest entirely on the small qualitative sample.","section":"Methods / Findings"},{"comment":"Discussion section: The leap from the single-company case to 'design considerations for inclusive deployment' in high-stakes HR environments (final paragraph of abstract and Discussion) is presented without an explicit limitations subsection addressing external validity, cross-site replication, or tests of whether the same positionalities predict adoption elsewhere.","section":"Discussion"}],"minor_comments":[{"comment":"The term 'work positionalities' is used repeatedly from the abstract onward but receives no explicit operational definition or citation to prior literature on the first use, which reduces clarity for readers outside the immediate subfield.","section":"Abstract / Introduction"}],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":{"model":"grok-4.3","summary":"We thank the referee for the detailed and constructive feedback. We address the two major comments point by point below, indicating the revisions we will make to strengthen the manuscript.","responses":[{"response":"We acknowledge the referee's observation. The search log data were used to guide participant recruitment and to provide background context on overall usage patterns during the transition, but no quantitative metrics or direct triangulation from the logs were reported to support the positionalities findings. The core claims rest on the survey and interview data. We will revise the Methods section to clarify the specific role of the log data, adjust the abstract and introduction to avoid overstating its contribution to the reported findings, and add any non-sensitive descriptive statistics from the logs if they can be included without privacy issues.","revision_made":"yes","referee_comment":"[Methods / Findings] Methods and Findings sections: The abstract and introduction state that search log data were analyzed alongside the survey and interviews, yet no quantitative results, metrics, or triangulation from the logs are reported to corroborate the positionalities (role/language/tenure) claims; the central findings therefore rest entirely on the small qualitative sample."},{"response":"We agree that an explicit limitations discussion is warranted. We will add a dedicated Limitations subsection that directly addresses the single-site case-study design, the modest sample (n=25 survey + 10 interviews), constraints on external validity, and the absence of cross-site replication or predictive testing. The design considerations will be reframed as case-derived insights intended to inform future work rather than as broadly generalizable prescriptions.","revision_made":"yes","referee_comment":"[Discussion] Discussion section: The leap from the single-company case to 'design considerations for inclusive deployment' in high-stakes HR environments (final paragraph of abstract and Discussion) is presented without an explicit limitations subsection addressing external validity, cross-site replication, or tests of whether the same positionalities predict adoption elsewhere."}],"tokens_in":1459,"tokens_out":422,"duration_ms":21774,"standing_objections":[]},"desk_editor":{"model":"grok-4.3","letter":"The main takeaway is that this paper supplies original observations from a live switch to a GenAI HR search tool in one multinational tech firm, using logs plus 25 survey responses and 10 interviews. That combination produces concrete points about how role, language, and tenure shaped whether the system fit employees' work, and how people built trust through source checks and colleague input.\n\nThe work does well by staying tied to actual deployment data rather than lab hypotheticals. The positionalities framing and the trust-calibration behaviors are drawn directly from the interviews and feel like useful additions to the acceptance literature.\n\nThe soft spots are the scale and the leap to recommendations. With only 25 survey responses and 10 interviews from a single company, the patterns cannot be treated as stable enough to ground design advice for other high-stakes HR settings. The abstract flags the logs but the summary gives no sign of quantitative triangulation that would strengthen the claims. The stress-test concern about untested generalizability holds up on the evidence provided.\n\nThis is for HCI and organizational researchers who want a documented case of workplace GenAI adoption. It deserves peer review because the data are original and the questions are relevant, even though the authors will need to narrow the scope of the conclusions or add more supporting analysis.","headline":"A grounded single-case study of GenAI rollout in one HR department that adds real transition data but rests its design recommendations on limited evidence.","tokens_in":2377,"tokens_out":331,"would_cite":false,"duration_ms":19727,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"grok-4.3","headline":"GenAI adoption in HR succeeds when the system's design assumptions align with employees' roles, spoken languages, and tenure.","keywords":["generative AI","workplace adoption","human resources","sociotechnical systems","trust in AI","organizational change","AI deployment","search systems"],"falsifier":"A study in another organization or with a larger workforce where GenAI adoption rates show no link to differences in role, language, or tenure.","tokens_in":2708,"feed_emoji":"","tokens_out":601,"duration_ms":22436,"temperature":0.7,"pith_summary":"The paper studies a multinational tech company replacing its legacy HR search system with a GenAI-supported one. Analysis of logs, a 25-person survey, and ten interviews shows uneven adoption driven by how well the tool's built-in assumptions match workers' specific positions. Trust formed when employees could check sources, compare outputs across systems, and consult colleagues or HR. The authors argue these sociotechnical conditions must be addressed for inclusive rollout in high-stakes settings. Their evidence points to treating company knowledge resources as part of the AI infrastructure itself.","feed_headline":"GenAI in HR gets used more when it matches roles and languages","feed_subtitle":"One company transition showed adoption hinged on design fit with employees' positions plus habits of verification and colleague checks.","key_machinery":"The fit between GenAI design assumptions and employees' work positionalities (role, spoken language, tenure)","core_discovery":"Adoption depended on the fit between the GenAI system's design assumptions and employees' work positionalities (role, spoken language, tenure). Employees built trust in GenAI answers through source-checking, comparison among systems, and seeking input from colleagues or HR when in doubt. The study supplies empirical evidence from a live organizational transition and converts the patterns into design considerations for inclusive deployment.","pith_inferences":["The same positional fit issues could appear in other knowledge-work settings where information quality affects decisions.","Targeted design adjustments for language or tenure groups might raise overall adoption without changing the underlying model.","Longitudinal tracking of the same employees after initial rollout could reveal whether trust-building habits persist or change."],"forward_implications":["Organizations must design GenAI systems to account for role- and context-sensitive benefits across different employee groups.","Treating organizational knowledge infrastructure as AI infrastructure improves accountability and usability of GenAI systems.","Search literacy, trust calibration, content quality, training, and guidance shape adoption alongside positional fit.","Inclusive deployment requires attention to situational fit rather than uniform rollout."],"fun_headline_variants":["GenAI in HR used more with role and language alignment","HR GenAI trust requires verification and peer checks","Employee roles shape GenAI adoption in live HR shift","Design GenAI HR systems for positional employee needs"],"cache_read_input_tokens":2112,"weakest_assumption_plain":"Observations from a single tech company's HR transition and a small sample of participants can inform broader GenAI deployment in other high-stakes environments.","fun_headline_variants_meta":{"raw":{"variants":["GenAI in HR used more with role and language alignment","HR GenAI trust requires verification and peer checks","Employee roles shape GenAI adoption in live HR shift","Design GenAI HR systems for positional employee needs"]},"model":"grok-4.3","cost_usd":0.007122,"raw_usage":{"total_tokens":3312,"prompt_tokens":711,"num_sources_used":0,"completion_tokens":60,"cost_in_usd_ticks":71224500,"prompt_tokens_details":{"text_tokens":711,"audio_tokens":0,"image_tokens":0,"cached_tokens":256},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":2541,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":711,"tokens_out":60,"duration_ms":21896,"temperature":1.0,"reasoning_tokens":2541,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-06-26T22:58:50.203592+00:00","model_set":{"reader":"grok-4.3"},"falsifier":"A study in another organization or with a larger workforce where GenAI adoption rates show no link to differences in role, language, or tenure.","supporting_citations":[],"review_version":1}