{"id":"c4bf00c5-57c3-47d3-9216-c93601d16d08","arxiv_id":"2505.02329","paper_version":3,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A multi-stakeholder interview study of US Fair Workweek scheduling shows that algorithmic management regulation is shaped by institutional constraints, software use in practice, enforcement mismatches, and software-specific concerns.","lead":"Researchers interviewed 38 people involved in workplace scheduling rules, software, and enforcement to find out what happens when labor laws are turned into software. They found that the success of algorithmic management regulation depends less on the software alone and more on how rules, software, and real workplaces interact.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Missing vendor/developer interviews leave a key institutional-constraints finding secondhand and untested.","rationale":"We read the paper in good faith as an exploratory, multi-stakeholder qualitative study whose central claim is explicitly hedged as a set of suggested findings. The four-factor framework is a synthesis of stakeholder perceptions, not a causal estimate, and the paper discloses its key limitation in Section 7: software vendors and developers were not directly interviewed. We agree with the reader that the most load-bearing assumption is that the 38 interviewees provide a sufficiently accurate and complete picture of the regulatory process. We sharpen that concern by focusing on the vendor gap as particularly consequential: factor (i) includes specific assertions about vendor motivations (e.g., insufficient financial incentives, lack of proactive guidance-seeking) that are made by regulators and attorneys, not by vendors themselves. These assertions are not incidental detail—they directly inform the paper's practical recommendations in Section 6. If vendor accounts diverge, those recommendations would need adjustment. However, the paper's cautious wording ('our findings suggest') and its explicit acknowledgment of the missing vendor perspective mean the current ACCEPT verdict remains appropriate; the concern is a reason to conduct follow-up work, not to reject the study as internally inconsistent or unsound. The proposed concrete test—direct vendor/developer interviews focused on the specific claims in Sections 5.1 and 5.4—would settle whether the secondhand attributions hold.","tokens_in":32315,"tokens_out":4721,"duration_ms":58897,"concrete_test":"Conduct follow-up interviews with 10–15 representatives from the scheduling-software vendors whose products appear in Table 8 (e.g., UKG/Kronos, Deputy, 7shifts, Homebase, Workday), using a protocol that asks specifically about (a) how they decide which jurisdiction-specific FWW features to implement, (b) their interactions with regulators during rulemaking, and (c) their views on the consent-UI examples in Section 5.4.2. If vendors' accounts of their constraints and incentives are consistent with the secondhand reports in Sections 5.1.2–5.1.3, the concern is resolved; if they materially diverge, factor (i) and the related design/policy implications in Section 6 require revision.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's factor (i) asserts that institutional constraints—including software vendors' insufficient financial incentives (Section 5.1.3) and their failure to seek guidance proactively (Section 5.1.2)—challenge the encoding of law into AM software. The evidence for these vendor-side claims comes exclusively from defense attorneys and regulators (D1–D4, R1–R8), not from vendors themselves. This is load-bearing because the proposed remedies in Section 6—boundary objects to align stakeholder incentives (6.3), disclosure-based approaches (6.4.2), and strategies for regulatory capacity (6.4.4)—depend on an accurate diagnosis of vendor constraints. If vendors' actual decision calculus is driven by legal ambiguity, the cost of jurisdiction-specific configuration, or employer demand rather than a lack of incentives, the paper's implications for 'aligning incentives' and designing boundary objects could be misdirected. The limitation is acknowledged in Section 7, but acknowledgment does not reduce the evidentiary weight here: a core element of the framework is supported only by secondhand attributions about an absent stakeholder group.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper presents a qualitative interview study (N=38) of stakeholders involved in the regulation of algorithmic management in the workplace, focused on U.S. Fair Workweek laws and shift-scheduling software. The authors interviewed regulators, corporate defense attorneys, worker advocates, scheduling managers, and workers, and supplemented this with a review of 38 scheduling-software product websites. Their central claim is that the efficacy of AM regulation is shaped by four factors: (i) institutional constraints in encoding law into software, (ii) on-the-ground software use that affects compliance, (iii) mismatches between software and regulatory contexts that hinder enforcement, and (iv) software-specific concerns introduced when software is used as a regulatory tool. The paper proposes a sociotechnical approach to AM regulation, arguing for multi-stakeholder engagement, boundary objects, and design- and policy-oriented recommendations for aligning software and law.","tokens_in":32446,"tokens_out":4632,"duration_ms":58131,"significance":"If the framework holds, the paper provides one of the first empirical, cross-stakeholder accounts of how workplace scheduling regulation actually operates in practice, spanning rule operationalization, software use, and enforcement. This is a meaningful contribution to HCI, software engineering, and regulatory studies, where much prior work is theoretical, experimental, or limited to a single stakeholder group. The paper's strengths include extensive participant quotes, detailed appendices documenting participant characteristics and the software review, full interview protocols, and consistently hedged language ('suggest') when generalizing from interview data. The multi-stakeholder design, including workers and managers alongside legal and regulatory actors, is a genuine advance over prior single-stakeholder studies. The design implications and future research agenda are useful even if the empirical base is not without limitations.","major_comments":[{"comment":"The findings about software vendors' behavior are supported only by secondhand accounts: §5.1.2 attributes vendors' failure to seek guidance proactively and their reliance on the wrong expertise to defense attorneys and regulators (D2, D3, R6), and §5.1.3 attributes the lack of financial incentives to D3, R8, R2, and D4, while Section 7 concedes that software vendors and developers were not directly interviewed. This gap is load-bearing because factor (i) and the remedies in Sections 6.3 and 6.4 (boundary objects, disclosure approaches, regulatory capacity) are predicated on an accurate diagnosis of vendor-side constraints; if vendor behavior is instead driven by legal ambiguity, the cost of jurisdiction-specific configuration, or employer demand, the recommendations could be misdirected. The authors should either re-scope these findings as stakeholders' perceptions of vendor constraints, add direct vendor/developer data or artifact-based triangulation, and clearly mark the recommendations as conditional on that evidentiary base.","section":"§5.1.2, §5.1.3, §7"},{"comment":"The central claim is stated as 'the efficacy of AM regulation is influenced by' the four factors, but the empirical basis is semi-structured interviews about experiences and beliefs, not measured compliance outcomes or observed enforcement results. This is a construct-level mismatch between the data and the claimed scope of the finding. I recommend consistently referring to 'stakeholders' perceived or reported challenges to efficacy' throughout the abstract, findings, and conclusion, and noting in Section 7 that actual regulatory effectiveness remains an open empirical question, so that the conclusions do not overstate what interview data can establish.","section":"§1, §5 opening, §8"}],"minor_comments":[{"comment":"The software name 'DeDoose' appears to be a typo for 'Dedoose'; please correct it.","section":"§3.2"},{"comment":"The quotation 'the exact opposite of what [FWW] law envisions' contains an awkward bracket; consider writing 'what the FWW law envisions' or 'what the Fair Workweek law envisions.'","section":"§5.2.2"},{"comment":"Figure 1, which is central to understanding the four-factor framework, is placed in Appendix A but referenced only as 'Figure 1 in Appendix A'; consider moving it into the main text near the opening of Section 5 or providing a more prominent cross-reference in the Findings section.","section":"§5 and Appendix A"},{"comment":"Capitalization of 'Fair Workweek Law' is inconsistent across the manuscript (e.g., 'FWW Laws' vs. 'FWW Law'); please unify the terminology.","section":"§4.1"}],"recommendation":"major_revision","confidential_remarks":"The main reason for major revision is the evidentiary gap regarding software vendors' perspectives, which is load-bearing for factor (i) and the associated design and policy recommendations. The paper is otherwise a solid qualitative contribution, and I would be willing to accept after the authors re-scope the vendor-related claims and adjust the central framing from 'efficacy' to 'perceived or experienced challenges to efficacy.'"},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"This is the first study I know that traces a full algorithmic-management regulation loop—rule operationalization, software use, enforcement—in a real implemented domain. It earns a serious referee. The four-factor framework is a useful map, and the paper is honestly hedged throughout.\n\nWhat's actually new: they interviewed 38 people across five stakeholder groups, including corporate defense attorneys and workers, who are usually missing from Fair Workweek compliance studies. The review of 38 scheduling product websites adds a nice empirical baseline. The appendices are detailed enough to follow the coding logic, and the quotes are abundant.\n\nSoft spots, in order of weight. The biggest is that the vendor-side findings—especially the \"lack of financial incentives\" claim in 5.1.3 and the \"vendors don't seek guidance\" observation—are based on regulators' and defense attorneys' accounts, not on vendor interviews. The limitation is acknowledged in Section 7, and the findings are framed as perceptions, so the central framework survives, but the design recommendations in Section 6 lean on that secondhand diagnosis more than they should. I'd want vendor interviews or at least a softer set of implications. Second, recruitment of workers and managers via subreddits and Facebook ads could skew toward people with strong experiences; the authors note this. Third, there is no measured compliance baseline, so everything about efficacy rests on perceptions. That is normal for qualitative work, but it means the paper should not be read as an evaluation of whether Fair Workweek actually works.\n\nThe stress-test concern about vendor interviews is partially right, but not fatal, because the paper's own claims are about stakeholder experiences, not vendor reality. On citation pattern: seems fine; self-citations are used for context, not load-bearing.\n\nWho this is for: HCI, CSCW, and FAccT readers studying algorithmic management, labor law compliance, or policy design. I'd take it to the reading group and cite it. Verdict: accept with minor revisions; the vendor gap should be discussed or the claims in Section 5.1.3 softened.","headline":"First full-loop empirical map of Fair Workweek regulation—solid, honestly hedged, with a real but acknowledged vendor-side blind spot.","tokens_in":32980,"tokens_out":2157,"would_cite":true,"duration_ms":26953,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper claims that regulating algorithmic management succeeds only when legal rules, software, and everyday workplace practice are aligned across the whole regulatory chain, from rulemaking to enforcement.","keywords":["algorithmic management","workplace scheduling","Fair Workweek laws","labor law compliance","sociotechnical systems","regulatory enforcement","worker consent","qualitative interviews"],"falsifier":"A direct interview and code-inspection study of scheduling-software vendors and developers would settle whether the institutional-constraints factor is accurate. If vendors can produce detailed legal requirement documents, dedicated compliance staff, and evidence that local-law customization pays off, the paper's first factor would be weakened; if they instead describe ambiguous guidance, weak incentives, and translation failures, it would be confirmed.","tokens_in":32107,"feed_emoji":"⚖️","tokens_out":8869,"duration_ms":101017,"temperature":0.7,"pith_summary":"Fair Workweek laws were meant to curb unpredictable scheduling, and scheduling software is now marketed as a compliance tool, but this study argues the laws' effectiveness depends on more than whether the software can compute a rule. Drawing on interviews with 38 regulators, defense attorneys, worker advocates, managers, and workers, the paper identifies four determinants of regulatory efficacy: institutional constraints on encoding law into software, on-the-ground use that shapes compliance, mismatches between software and regulatory contexts that obstruct enforcement, and new concerns software raises when it mediates regulation. The consequence is that coding a legal requirement into a product is not enough; the organizations, routines, and enforcement capacities around the software decide whether the law protects workers. The paper concludes that algorithmic-management regulation needs a sociotechnical approach and offers design and policy directions for aligning software with the law.","feed_headline":"Scheduling software alone can't enforce fair-workweek laws","feed_subtitle":"Interviews with 38 stakeholders show compliance fails across rulemaking, software use, and enforcement.","key_machinery":"The central object is the AM regulatory process, a three-stage pipeline: rule operationalization, software use, and enforcement. The paper treats this pipeline as a single sociotechnical system in which each stage's outputs become the next stage's inputs, so guidance becomes software logic, software logic becomes workplace records, and records become the basis for investigations. Misalignment at any stage propagates forward, which is why regulatory efficacy cannot be evaluated by looking at software features alone. The paper also introduces boundary objects, shared instruments such as personas, flowcharts, pseudocode, and data probes, as the proposed mechanism for keeping stakeholders aligned across the pipeline.","core_discovery":"On the paper's own terms, the central discovery is an empirical map of why algorithmic-management regulation underdelivers, drawn from the one AM domain with implemented laws: workplace scheduling. At the rule-operationalization stage, regulatory guidance leaves ambiguous use cases unresolved, legal-technical collaboration is adversarial and error-prone, and vendors lack financial incentives to tailor products to local jurisdictions, so provisions like access-to-hours and good-faith estimates get encoded unevenly or not at all. At the software-use stage, operational demands push managers and workers around the software, so consent is collected after a shift, requests are made in person, and managers override recorded reasons, meaning the data the software generates does not reliably reflect whether the law was followed. At the enforcement stage, rules lack measurable thresholds, data is inconsistent and scattered, procedural evidence such as proof of a one-on-one consent conversation is missing, and agencies lack data-analysis capacity, so violations become hard to prove. The paper also finds that software itself introduces concerns: interaction designs that nudge workers to waive premium pay, checkbox consent that workers later dispute, and opaque vendors that either overpromise compliance or silently omit it. These four factors, the paper argues, jointly determine whether Fair Workweek regulation actually protects workers.","pith_inferences":["A testable extension beyond this paper's scheduling context: if the four-factor framework generalizes, algorithmic-hiring and platform-work enforcement should show analogous breakdowns at the rule operationalization, use, and enforcement stages, and comparing those regimes would sharpen or revise the account.","The paper's account implies that compliance is not a property of a software product but an outcome of relationships among vendor, employer, manager, worker, and regulator; a procurement or audit standard built on this view would evaluate deployment context, not just feature checklists.","Directly interviewing software vendors and developers, which the paper identifies as future work, could reveal whether the institutional constraints are best explained by legal ambiguity, weak financial incentives, or technical debt and code architecture; the current findings rely on secondhand reports from regulators and attorneys.","The boundary-object recommendation is testable: a co-design workshop with personas, flowcharts, and data probes could measure whether these artifacts reduce the lost-in-translation errors documented in the paper, and whether they change how vendors encode ambiguous provisions."],"forward_implications":["If the four-factor account is right, regulations that only specify what software must do, without addressing vendor incentives, legal-technical collaboration, and agency capacity, will leave enforcement gaps that surface later as unprovable violations.","Compliance audits should examine software interfaces and interaction flows, not just the data they produce, because design choices like default waivers and checkbox consent can manufacture records that look lawful while workers are harmed.","Enforcement agencies need data-analysis staffing, tooling, and standardized record formats, otherwise even well-encoded rules cannot be converted into legal findings from the records employers provide.","The appropriate level of automation in compliance software is a policy decision with trade-offs: full automation can prevent violations but hides errors and lacks context, while human discretion is flexible but prone to coercion and error.","Procedurally valid compliance requires capturing evidence of process, such as timestamps, one-on-one conversations, and informed consent, not just outcomes; data contextualization and anonymous worker reporting channels are concrete design directions."],"supporting_citations":[{"why":"Documents the ambiguity problem in translating legal texts into software logic, which the paper finds persists in Fair Workweek rulemaking.","marker":"[77]"},{"why":"Supplies the requirements-engineering challenges that arise when encoding legal requirements, serving as the baseline the paper extends to scheduling software.","marker":"[88]"},{"why":"Proposes a methodology for encoding legislation; the paper contrasts this with real-world implementations that lack usage data.","marker":"[118]"},{"why":"Presents an experimental study of law and computer science students translating bankruptcy code, the comparison point for real-world legal-software collaboration.","marker":"[32]"},{"why":"Evaluates employer compliance with Seattle's secure scheduling ordinance, providing the empirical context of Fair Workweek outcomes that the paper builds on.","marker":"[47]"},{"why":"Reports challenges regulators faced in supporting employers with Fair Workweek adoption, which the paper extends to enforcement and defense-side perspectives.","marker":"[75]"},{"why":"Interviews algorithmic-hiring auditors, representing the single-stage regulatory analysis that the paper aims to broaden.","marker":"[45]"},{"why":"Defines boundary objects as shared instruments for cross-group coordination, the mechanism the paper proposes for aligning regulatory stakeholders.","marker":"[16]"},{"why":"Originates the boundary-object concept, supporting the paper's argument for personas, flowcharts, pseudocode, and data probes.","marker":"[109]"},{"why":"Shows how professionals deviate from intended algorithm use in practice, supporting the paper's finding on on-the-ground software use.","marker":"[20]"}],"fun_headline_variants":["Why fair-workweek laws fail in practice: 38 stakeholders weigh in","Algorithmic scheduling: the law and software are out of sync","Study: software alone can't make scheduling laws work","Regulating AI scheduling: where the rules break down","Fair workweek enforcement gaps revealed in multi-stakeholder study"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The four-factor account rests on the assumption that 38 self-selected interviewees, none of them software vendors or developers, give a complete enough picture of how scheduling software is built and used; if vendor-side constraints and motives differ substantially from these secondhand accounts, the institutional-constraints finding collapses.","fun_headline_variants_meta":{"raw":{"variants":["Why fair-workweek laws fail in practice: 38 stakeholders weigh in","Algorithmic scheduling: the law and software are out of sync","Study: software alone can't make scheduling laws work","Regulating AI scheduling: where the rules break down","Fair workweek enforcement gaps revealed in multi-stakeholder study"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000708,"raw_usage":{"total_tokens":3208,"prompt_tokens":981,"completion_tokens":2227,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":597,"completion_tokens_details":{"reasoning_tokens":2143}},"tokens_in":597,"tokens_out":2227,"duration_ms":17498,"temperature":1.0,"reasoning_tokens":2143,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T00:54:24.934556+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A direct interview and code-inspection study of scheduling-software vendors and developers would settle whether the institutional-constraints factor is accurate. If vendors can produce detailed legal requirement documents, dedicated compliance staff, and evidence that local-law customization pays off, the paper's first factor would be weakened; if they instead describe ambiguous guidance, weak incentives, and translation failures, it would be confirmed.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the requirements-engineering challenges that arise when encoding legal requirements, serving as the baseline the paper extends to scheduling software."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Proposes a methodology for encoding legislation; the paper contrasts this with real-world implementations that lack usage data."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Shows how professionals deviate from intended algorithm use in practice, supporting the paper's finding on on-the-ground software use."}],"review_version":1}