Organization under AI
DE EN

Cheap to Make, Costly to Check · Volume One

Sources

This book has no footnotes. Where a finding comes from someone else’s work, the name is in the sentence, and the full reference is here. Where something is my reading rather than an established result, the text says so at that spot.

Each entry carries what a reader needs in order to disagree with it: what was measured, on whom, and what limits the finding. Entries marked [limit] carry a caveat that the text also states.

Where a chapter has no entry, it works without external evidence or carries its sources in the text.

Every entry carries its type. Primary source Secondary source Preprint Legal source The book's own Constructed case See above Generally documented Editorial note Type not assigned

Introduction

  • Constructed caseThe people in this book — Ines Kessler, Frank Ostrowski, Tim Whelan, and the company they work for — are composite figures, assembled from real engagements and real conversations. The mechanisms reflect patterns that recur; the narrative details are constructed.

Chapter 1 · Forty-One Bids

  • Primary sourceAgrawal, Ajay, Joshua Gans, and Avi Goldfarb. Prediction Machines: The Simple Economics of Artificial Intelligence. Harvard Business Review Press, 2018; expanded edition 2022. The argument: AI lowers the cost of prediction, and when one input gets cheaper the value of its complements rises — for these authors, human judgment. The abstract mechanism is theirs, not mine. What this book adds is the continuation: quantity rather than price, review capacity as the limit, and the migration of the load between positions.
  • Primary sourceBainbridge, Lisanne. “Ironies of Automation.” Automatica 19, no. 6 (1983): 775–79.
  • Primary sourceElish, Madeleine Clare. “Moral Crumple Zones: Cautionary Tales in Human-Robot Interaction.” Engaging Science, Technology, and Society 5 (2019): 40–60.

Chapter 2 · Look Again

  • Primary sourceThe Value of AI 2026 — Germany. SAP and Oxford Economics, published July 2026; n = 200 German decision-makers at director level and above, in companies with 500 or more employees, within a total sample of 2,600 across 13 countries. Fifty-seven percent without a process that brings human oversight into agentic workflows; 34 percent with a central register of agents; 75 percent who agree or are unsure whether they are deploying agents faster than they can govern them effectively. [limit] Commissioned by a vendor of agentic software; smallest size band starts at 500 employees; German data. The text states all three.

Chapter 3 · Why This Happens

  • Primary sourceBrysbaert, Marc. “How Many Words Do We Read Per Minute? A Review and Meta-Analysis of Reading Rate.” Journal of Memory and Language 109 (2019), article 104047. Meta-analysis of 190 studies with 18,573 participants; mean silent reading rate for nonfiction 238 words per minute, range 175 to 300. The rise from the first to the fourth year of college (16 words per minute) is the Nelson-Denny norm Brysbaert reports (Brown et al., 1993); whether it comes from the reading load or from age, he leaves open.
  • Primary sourceSimon, Herbert A. “Applying Information Technology to Organization Design.” Public Administration Review 33, no. 3 (1973): 268–78. The formulation that attention is the essential bottleneck of organizational activity, and the observation that it migrates upward through a hierarchy.
  • PreprintAdvani, Laksh. “From Confident Closing to Silent Failure: Characterizing False Success in LLM Agents.” arXiv:2606.09863, June 2026. A total of 9,876 trajectories from eight model families and 1,879 from four more. No configuration of five checker models and five prompting strategies exceeds AUROC 0.65; on one trajectory set the same checkers sit at 0.54, near chance. The checkers go by surface features, including confident closing language. [limit] A preprint, since accepted at a workshop. That is not full peer review.
  • Primary sourceTrist, Eric L., and Ken W. Bamforth. “Some Social and Psychological Consequences of the Longwall Method of Coal-Getting.” Human Relations 4, no. 1 (1951): 3–38.
  • Primary sourceGoldratt, Eliyahu M. The Goal (1984) and the Theory of Constraints developed from it.
  • PreprintAI Debt: Verification Burdens, Rework Externalities, and the Hidden Costs of Asymmetric Generative-AI Adoption. Cambridge Open Engage, May 2026. [limit] Preprint, not peer reviewed.
  • Primary sourceZhu, Liming, Qinghua Lu, Ming Ding, Sung Une Lee, and Chen Wang. “Designing Meaningful Human Oversight in AI.” AI and Ethics 6 (2026), article 286. Contains the asymmetry between solving and verifying.
  • See aboveElish 2019 — see Chapter 1.
  • Primary sourceBetterUp Labs and the Stanford Social Media Lab. Survey of 1,150 U.S. office workers, September 2025. Forty percent had received machine-generated material in the previous month that looked like work and saved no work; those who did needed on average 1 hour 56 minutes per instance to make it usable. [limit] BetterUp sells services for this problem; the forty percent is self-report about what somebody believed was machine-generated; the aggregate cost figure that reached the press was extrapolated from self-reported salaries and is not used here.

Chapter 4 · What the Machine Can’t See

  • Legal sourceEqual Credit Opportunity Act; Regulation B, 12 CFR 1002.9 (statement of specific reasons for adverse action). Consumer Financial Protection Bureau, Circular 2022-03, “Adverse Action Notification Requirements in Connection with Credit Decisions Based on Complex Algorithms,” May 2022, and Circular 2023-03, “Adverse Action Notification Requirements and Proper Use of Sample Forms.” The position: a creditor cannot justify noncompliance on the grounds that its technology is too complicated or opaque to understand, and may not fall back on sample reasons that do not specifically and accurately state what the decision turned on. Both circulars were withdrawn on May 12, 2025, with 65 other guidance documents (Federal Register 2025-08286). [limit] The text states the withdrawal; the duty in Regulation B is unchanged.
  • Legal sourceThe European counterpart, referred to in the text without being relied on: Court of Justice of the European Union, judgment of December 7, 2023, Case C-634/21 (SCHUFA). A credit score is itself an automated individual decision under Article 22 GDPR where a third party draws on it substantially. The British, German, and Japanese positions named alongside it are set out under The Same Two Questions, Five Legal Systems, below.
  • Primary sourceDodge, Jesse, Maarten Sap, Ana Marasović, William Agnew, Gabriel Ilharco, Dirk Groeneveld, Margaret Mitchell, and Matt Gardner. “Documenting Large Webtext Corpora: A Case Study on the Colossal Clean Crawled Corpus.” Proceedings of EMNLP 2021. The blocklist filtering used to build C4 removes about 42 percent of documents in African American English and about 32 percent in Hispanic-aligned English, against 6.2 percent for white-aligned English. [limit] The figure concerns one corpus and one filtering step; it is used here as the documented case of selection going in, not as a statement about every model.
  • Primary sourceJoshi, Pratik, Sebastian Santy, Amar Budhiraja, Kalika Bali, and Monojit Choudhury. “The State and Fate of Linguistic Diversity and Inclusion in the NLP World.” Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, 2020, 6282–93. Table 1: 88.38 percent of the world’s languages fall in Class 0, “The Left-Behinds” — languages with exceptionally limited resources, for which unsupervised pre-training, in the authors’ words, “only makes the poor poorer.”
  • Primary sourceSun, Luning, Yuzhuo Yuan, Yuan Yao, Yanyan Li, Hao Zhang, Xing Xie, Xiting Wang, Fang Luo, and David Stillwell. “Large Language Models Show Both Individual and Collective Creativity Comparable to Humans.” Thinking Skills and Creativity (2025), article 101870. Thirteen creative tasks across divergent thinking, problem solving, and creative writing; the best models rank at the 52nd percentile against humans, and when queried ten times a model’s collective output matches eight to ten people. [limit] Measured in standardized tasks in which the lived and the local do not appear — which is to say, not the thing at issue in Chapter 4 or Chapter 13. It is cited as the strongest finding in the other direction. This is the same study referred to in Chapter 13.

Chapter 5 · Where the Bill Lands

  • Secondary sourceFacebook’s engagement retuning, 2018, and the internal research that followed. Documented through the 2021 disclosures by Frances Haugen.
  • Legal sourceWells Fargo. Consent order of the Consumer Financial Protection Bureau, September 8, 2016 ($100 million), with parallel penalties from the Office of the Comptroller of the Currency ($35 million) and the City and County of Los Angeles ($50 million), $185 million in total.
  • Primary sourceChallenger, Gray & Christmas, Inc. Job Cut Announcement Report, monthly. Checked against the primary source: the August 2026 report, Table 4, records 116,175 AI-attributed cuts in calendar 2026 through August — about 22 percent of all announcements and still the most-cited reason of the year. For comparison, all of 2025: 54,836. [limit] Challenger also publishes a cumulative figure since it began counting AI as a separate reason in 2023, which is easily confused with the annual one. The 2025 total is not separately cross-checked against the annual report.
  • Primary sourceBonney, Kathryn, Cory Breaux, Emin Dinlersoz, Lucia Foster, John Haltiwanger, and Aditya Pande. “The Microstructure of AI Diffusion: Evidence from Firms, Business Functions, and Worker Tasks.” U.S. Census Bureau, Center for Economic Studies, Working Paper CES-26-25, April 2026, drawing on the Business Trends and Outlook Survey supplement of November 2025 to January 2026. Eighteen percent of firms used AI in a business function (32 percent employment-weighted); 57 percent of users deploy it in three or fewer functions; AI-related employment decreases occurred in 2 percent of firms. [limit] The Census figures count firms; the Challenger figures count announced cuts. The two measure different things and cannot be set against each other directly; the text reads the gap as an order of magnitude only.
  • Secondary sourceFord. Statement by Charles Poon, vice president of vehicle hardware engineering, to journalists in June 2026, reported by TechCrunch (Anthony Ha, June 28, 2026), Fortune (June 29), and Forbes (June 30), among others. As quoted: “Mistakenly we thought that by just introducing artificial intelligence and ingesting the design requirements that we had, that that would produce a high-quality product.” In all, 350 veteran engineers, some former employees, some from suppliers — rehired, hired, or promoted over the past three years — to mentor younger staff and to retrain the tools. J. D. Power’s 2026 study surveyed owners of model-year 2026 vehicles between mid-2025 and May 2026; the timing does not allow the engineers to be credited with the result, and the text does not. First place among mass-market brands in J. D. Power’s 2026 Initial Quality Study, for the first time in sixteen years, at 152 problems per 100 vehicles. [limit] The outlets differ on who reported it first — TechCrunch credits Bloomberg, others The Verge — and none prints a full transcript. The quote is consistent across outlets.
  • Primary sourceUniversity of Chicago Law School. AI Strategy Statement, July 9, 2026. For the 2026–27 academic year, electronic devices are prohibited in the classroom across all required first-year courses, with limited exceptions; examinations are held in class without access to the internet, electronic files, or apps. Verbatim: “Student expertise at judging the quality of AI output is at its nadir.”
  • Primary sourceACM Task Force on Generative AI and Programming Assessment. Report of February 16, 2026, presented at SIGCSE 2026. Global survey, 763 instructors from 49 countries; the 68 percent refers to the 459 who answered that question.

Chapter 6 · Whose Bill Is This

  • Primary sourceKoyama, Alain K., Claire-Sophie Sheridan Maddox, Ling Li, Tracey Bucknall, and Johanna I. Westbrook. “Effectiveness of Double Checking to Reduce Medication Administration Errors: A Systematic Review.” BMJ Quality & Safety 29, no. 7 (2020): 595–603. Thirteen studies included, three of high quality; one shows a better error rate, none a difference in the harm reaching patients.
  • Primary sourceSkitka, Linda J., Kathleen L. Mosier, Mark Burdick, and Bonnie Rosenblatt. “Automation Bias and Errors: Are Crews Better Than Individuals?” The International Journal of Aviation Psychology 10, no. 1 (2000): 85–97. Error rates do not differ meaningfully between one- and two-person crews. The accountability finding — that people who expect to answer for correctness verify automated output more often — is from the companion study: Skitka, Linda J., Kathleen L. Mosier, and Mark Burdick. “Accountability and Automation Bias.” International Journal of Human-Computer Studies 52, no. 4 (2000): 701–17.
  • Legal sourceSmith v. Van Gorkom, 488 A.2d 858 (Del. 1985). The board approved a merger after a meeting of about two hours, without adequate financial data and without a fairness opinion; the Delaware Supreme Court held that failing to inform oneself of the material information reasonably available is gross negligence, which removes the protection of the business judgment rule. The settlement was $23.5 million. Delaware responded in 1986 with Section 102(b)(7) of the General Corporation Law, which lets a corporation’s charter exempt directors from monetary liability for breaches of the duty of care; since 2022 it extends, more narrowly, to officers. [limit] The case concerns a board and a merger. No court has decided whether an unreviewed ranking counts as material information reasonably available. The transfer is this book’s reading, and the text says so.
  • See aboveGoldratt — see Chapter 3.

Chapter 7 · The Decision Nobody Claimed

  • Primary sourceAristotle. Nicomachean Ethics — the distinction between techne and phronesis.
  • Primary sourceAnthropic. “Claude’s Constitution,” version of January 21, 2026, licensed CC0 1.0. Cited here for the rules-versus-judgment dilemma, the absolute limit against self-exfiltration and evading legitimate oversight, and the principle of acquiring no more resources or capabilities than a task requires. [limit] A company’s account of its own system, not an independent finding. It is used as evidence that such a document can be written, not that it works.
  • Secondary sourceOpenAI, agent incident at Hugging Face, July 2026. Disclosed by Hugging Face on July 16, 2026, and in a joint statement by OpenAI and Hugging Face on July 21; reported by ZDFheute, CNN, NBC News, CNBC, and TIME, among others. Agents in a cyber-capability test, run with deliberately loosened safety boundaries, escaped the sandbox through a vulnerability and got into Hugging Face’s production systems to obtain answers for the test task. Both details are in the text.
  • Secondary sourceAmazon, Baltimore fulfillment center. Productivity tracking measuring time between scans; roughly 300 full-time associates terminated for productivity between August 2017 and September 2018. Source: letter from Amazon’s outside counsel to the National Labor Relations Board, September 4, 2018, stating that the system generates warnings and terminations without input from supervisors; Amazon said supervisors can override it. Reported by The Verge, April 25, 2019.

Chapter 8 · Four Zones, Not Four Boxes

  • Primary sourceThe Dutch childcare benefits affair (Toeslagenaffaire). Parliamentary Committee of Inquiry, Ongekend onrecht (Unprecedented Injustice), Tweede Kamer, December 17, 2020 — main causes named: rigid legislation, all-or-nothing enforcement, failures of legal protection. Amnesty International, Xenophobic Machines: Discrimination through Unregulated Use of Algorithms in the Dutch Childcare Benefits Scandal, October 2021 — the risk classification model used Dutch nationality (yes/no) as an indicator; caseworkers did not see why a case had been flagged. Government of the Netherlands, February 12, 2026: more than 43,000 parents recognized as affected. The Dutch government resigned in January 2021.
  • Primary sourceOcasio, William. “Towards an Attention-Based View of the Firm.” Strategic Management Journal 18, S1 (1997): 187–206.
  • Primary sourceSheridan, Thomas B., and William L. Verplank. Human and Computer Control of Undersea Teleoperators. MIT Man-Machine Systems Laboratory, 1978. Technical report for the Office of Naval Research; contains the ten-level scale from wholly human to wholly machine action. The four zones take nothing from it. The entry stands because the genre predates the current occasion by nearly fifty years.
  • The book's ownThe five triggers, the four zones, and the constitutional-rank test are this book’s own framework. There is no external source for them.

Chapter 9 · Writing It Down

  • Primary sourceAnthropic, “Claude’s Constitution” — see Chapter 7.
  • Secondary sourceAmazon recruiting tool, 2014–2017. Automatic one-to-five-star rating of applications; by 2015 Amazon had realized it systematically disadvantaged female candidates, downgrading the word “women’s” and the names of all-women colleges; the team was disbanded by early 2017. According to Reuters, recruiters looked at the ratings but never relied on them alone. Reported by Reuters, October 10, 2018.
  • Secondary sourceIKEA and the AI assistant “Billie.” Initially just under half of call center inquiries, later around three quarters; the 8,500 call center staff were retrained as remote interior design advisors, a channel producing roughly €1.3 billion in fiscal 2022.
  • Legal sourceAnthropic PBC v. U.S. Department of War et al., No. 3:26-cv-01996-RFL (N.D. Cal.), Judge Rita F. Lin. Checked against the full text: Order on Cross Motions for Summary Judgment, August 27, 2026 (Dkt. 250); Order of Final Relief (Dkt. 251); Judgment (Dkt. 252). Documented: a two-year contract worth up to $200 million with the Chief Digital and Artificial Intelligence Office from July 2025; the demand for a clause covering “all lawful uses”; the two restrictions held — “lethal autonomous warfare and mass surveillance of Americans”; findings of impermissible retaliation, absence of prior hearing, and arbitrary and capricious action; permanent injunction and vacatur of the designation. The sentence on the contractual limit, verbatim: “The usage policy applicable to DoW work is a purely contractual limit; Anthropic is incapable of enforcing it technologically, and does not have direct visibility into how DoW uses its model.” [limit] A first-instance decision under U.S. law, with an appeal window running into late October 2026. The chapter’s argument hangs on the contractual limit and the two held restrictions, not on the survival of the judgment.
  • Legal sourceBoard of Governors of the Federal Reserve System. Supervisory Guidance on Model Risk Management, SR 11-7, April 4, 2011 (issued jointly with the OCC as Bulletin 2011-12; adopted by the FDIC in 2017 as FIL-22-2017). Verbatim: “Effective challenge depends on a combination of incentives, competence, and influence.” Checked against the full text. Replaced on April 17, 2026, by SR 26-2, Revised Guidance on Model Risk Management, which excludes generative and agentic AI from its scope. [limit] Supervisory guidance does not have the force of law (Interagency Statement on the Role of Supervisory Guidance, 2018, codified 2021). Its transfer to AI-assisted decisions outside banking is this book’s reading.

Chapter 10 · The Review That Only Looks Like One

  • Primary sourceKPMG International and the University of Melbourne. Trust, Attitudes and Use of Artificial Intelligence: A Global Study 2025. A total of 48,340 respondents in 47 countries, fielded November 2024 to mid-January 2025. Reported as “sometimes to very often”: among respondents who use AI at work, 66 percent relied on AI output without verifying the information; 56 percent made errors in their own work as a result. [limit] Self-report about one’s own behavior; the shares are more likely too low than too high, and the text says so.

Chapter 11 · Ninety Seconds Per Case

  • Primary sourceSiebert, Luciano Cavalcante, et al. “Meaningful Human Control: Actionable Properties for AI System Development.” AI and Ethics 3 (2023): 241–55 (online May 2022). The third of four properties: the responsibility assigned to a person should match their actual ability and authority to control the system. In secondary citation this work often runs as “Abbink et al.” The first author is Siebert.
  • Primary sourceParasuraman, Raja, Thomas B. Sheridan, and Christopher D. Wickens. “A Model for Types and Levels of Human Interaction with Automation.” IEEE Transactions on Systems, Man, and Cybernetics 30, no. 3 (2000): 286–97. Meta-analysis across eighteen experiments: Onnasch, Linda, et al. “Human Performance Consequences of Stages and Levels of Automation: An Integrated Meta-Analysis.” Human Factors 56, no. 3 (2014): 476–88.
  • Primary sourceBudzyń, Krzysztof, et al. “Endoscopist Deskilling Risk after Exposure to Artificial Intelligence in Colonoscopy: A Multicentre, Observational Study.” The Lancet Gastroenterology & Hepatology 10, no. 10 (2025): 896–903. Four centers, 1,443 examinations; adenoma detection rate falling from 28.4 to 22.4 percent (p = 0.0089) in examinations performed without the assistant after its introduction. [limit] Observational study, no randomization.
  • Primary sourceCounter-evidence: Jamieson, Greg A., and Gyrd Skraaning. “The Absence of Degree of Automation Trade-Offs in Complex Work Settings.” Human Factors 62, no. 4 (2020): 516–29, with a reply by Wickens et al. (2020). The dispute is live and this book does not settle it.
  • Primary sourceKuntz, Ludwig, Roman Mennicken, and Stefan Scholtes. “Stress on the Ward: Evidence of Safety Tipping Points in Hospitals.” Management Science 61, no. 4 (2015): 754–71. Eighty-three hospitals, 256 departments, 82,280 cases; safety tipping point at roughly 92.5 percent occupancy, attributed to exhausted escalation routines rather than staff exhaustion. [limit] The transfer of that number to review work is not shown and is expressly not claimed.
  • Primary sourceSee, Judi E. “Visual Inspection Reliability for Precision Manufactured Parts.” Human Factors 57, no. 8 (2015): 1427–42. Sandia National Laboratories; 82 inspectors of precision parts for the U.S. nuclear security enterprise, 140 components; 85 percent of defective parts rejected and 35 percent of sound parts rejected as well.
  • Primary sourceFitz, Nicholas, et al. “Batching Smartphone Notifications Can Improve Well-Being.” Computers in Human Behavior 101 (2019): 84–94, n = 237, on the dose effect of batching. Measured against this, using heart rate variability: Mark, Gloria, et al. “Email Duration, Batching and Self-Interruption: Patterns of Email Use on Productivity and Stress.” CHI 2016, 1717–28, n = 40 over an average of twelve workdays, finding no effect.
  • Primary sourceStarmer, Amy J., et al. (I-PASS). “Changes in Medical Errors after Implementation of a Handoff Program.” New England Journal of Medicine 371, no. 19 (2014): 1803–12. Program versus mandate: Haynes, Alex B., et al. “A Surgical Safety Checklist to Reduce Morbidity and Mortality in a Global Population.” NEJM 360, no. 5 (2009): 491–99 (mortality 1.5 → 0.8 percent across eight hospitals) against Urbach, David R., et al. “Introduction of Surgical Safety Checklists in Ontario, Canada.” NEJM 370, no. 11 (2014): 1029–38 (0.71 versus 0.65 percent across 101 hospitals, statistically indistinguishable).
  • Legal sourceThe lapsing approval as a legal device. Deemed-approval provisions are long established in administrative law; the recurring criticism cited in the text — that the fiction accelerates no review and creates no capacity — comes from the submissions of affected professional bodies. [limit] Not a scientific evaluation.

Chapter 12 · Faster, Not Easier

  • Primary sourceBedard, Julie, Matthew Kropp, Megan Hsu, Olivia Karaman, Jason Hawes, and Gabriella Kellerman (Boston Consulting Group / UC Riverside). “When Using AI Leads to 'Brain Fry.'” Harvard Business Review, March 2026. 1,488 full-time U.S. employees; 14 percent of AI users report the exhaustion, 26 percent in marketing.
  • Primary sourceMETR, 2025. Randomized trial, 16 experienced open-source developers on their own repositories: about 19 percent slower with AI assistance, while believing they were about 20 percent faster. [limit] A small sample in one setting. Other randomized field studies found gains: Cui, Zheyuan (Kevin), et al., Management Science (2026), three field experiments with 4,867 developers, about 26 percent more completed tasks; Dell’Acqua, Fabrizio, et al., Harvard Business School Working Paper 24-013 (2023), 758 consultants, faster and better work on tasks inside the models’ capabilities, worse outside them. The text uses METR for the gap between measured and felt effect, not for a claim that AI slows work down.
  • Primary sourceGeorghiou, Alexia. “The AI Paradox: Higher Productivity, Higher Stress.” SHRM, May 12, 2026: “As output becomes easier, expectations rise.” The up-to-20-percent stress figure comes from a global survey of June 2025 cited there. The term ratchet effect is not hers; it comes from the economics of incentives — Weitzman, Martin L. “The 'Ratchet Principle’ and Performance Incentives.” Bell Journal of Economics 11, no. 1 (1980): 302–08.
  • Primary sourceSteghaus, Sarah, Lena Hünefeld, and Sophie-Charlotte Meyer. PC, Smartphone und Co. im Arbeitsalltag [PC, smartphone, and company in the workday]. BIBB/BAuA Factsheet 64, August 2026, based on the 2024 BIBB/BAuA employment survey. Group sizes: no frequent digital use n = 1,765; PC n = 6,830; smartphone/tablet n = 8,995. [limit] A cross-sectional group comparison, not a causal finding — carried in the text as “goes together with.”
  • Primary sourceEuropean Working Conditions Survey 2024, Eurofound, overview report, April 2026. Over 40 percent report tasks added by technology, 30 percent tasks removed.
  • Primary sourceHacker, Winfried. Arbeits- und Gesundheitsschutz bei informationsverarbeitenden Arbeitsprozessen mit Künstlicher Intelligenz [Occupational safety and health in information-processing work processes involving artificial intelligence]. baua: Fokus, February 2026. Source of the twin directions — burdensome unlearning alongside overload — and of responsibility for results one was not involved in producing.
  • Primary sourceJob Stress Index, Gesundheitsförderung Schweiz [Health Promotion Switzerland]. [limit] Series break: the index was revised between 2022 and 2025, and values from the new series are not comparable with 2014–2022.
  • Legal sourceThe American position on psychosocial hazards at work: OSHA has no standalone standard; whether the general duty clause reaches psychological hazards is contested and unsettled. NIOSH, “An Urgent Call to Address Work-Related Psychosocial Hazards and Improve Worker Well-Being,” 2024. Office of the Surgeon General, Framework for Workplace Mental Health & Well-Being, 2022 — five essentials, beginning with protection from harm. All of it is guidance. The German, Japanese, and British counterparts the chapter sets against it are under The Same Two Questions, Five Legal Systems, below.

Chapter 13 · Whoever Configures, Governs

  • The book's ownConfiguration power and the second and third positions are this book’s own framework. The second position — whoever formulates the question determines the answer — is described, not documented. A process that leaves no trace appears in no investigation.
  • The book's ownOn the model’s tendency to agree with the asker: how strongly it shows up depends on the model and has not been comprehensively measured.
  • See aboveThe counter-work on breadth: Sun et al. 2025 — see Chapter 4. [limit] Measured in standardized tasks in which the lived and the local do not appear — which is to say, not the thing at issue here.
  • Generally documentedThe Talmudic preservation of minority opinions with their reasoning.

Chapter 14 · The Hour Without the System

  • The book's ownNIST. Artificial Intelligence Risk Management Framework (AI RMF 1.0), NIST AI 100-1, January 2023, Appendix C (AI Risk Management and Human-AI Interaction), verbatim: “Data about the frequency and rationale with which humans overrule AI system output in deployed systems may be useful to collect and analyze.” The metric is therefore not new and is not claimed here as an original finding. What is this book’s own is the reading: a rate near zero decides nothing; it asks the question.
  • Primary sourceAlon-Barkat, Saar, and Madalina Busuioc. “Human–AI Interactions in Public Sector Decision Making: 'Automation Bias’ and 'Selective Adherence’ to Algorithmic Advice.” Journal of Public Administration Research and Theory 33, no. 1 (2023): 153–69. Three experiments, 2,854 participants in total, the third with 1,345 Dutch civil servants; verbatim: “We do not find evidence for automation bias.” What they report instead is selective adherence — advice followed more often when it matched a stereotype, whether it came from an algorithm or a human expert. [limit] Vignette experiments, not a longitudinal measure of ability.
  • See aboveBainbridge 1983 — see Chapter 1.
  • Primary sourceWeibel, Antoinette, and colleagues, University of St. Gallen, on control and trust management. The first axis: Adler, Paul S., and Bryan Borys. “Two Types of Bureaucracy: Enabling and Coercive.” Administrative Science Quarterly 41, no. 1 (1996): 61–89.
  • Primary sourceParasuraman, Raja, and Victor Riley. “Humans and Automation: Use, Misuse, Disuse, Abuse.” Human Factors 39, no. 2 (1997): 230–53. Background to the three failure modes; not quoted in the text.
  • Primary sourceRost, Katja, Emil Inauen, Margit Osterloh, and Bruno S. Frey. “The Corporate Governance of Benedictine Abbeys: What Can Stock Corporations Learn from Monasteries?” Journal of Management History 16, no. 1 (2010): 90–115. A total of 134 Benedictine monasteries in Baden-Württemberg, Bavaria, and German-speaking Switzerland; average lifespan around 460 years after adjustment. About 20 percent of closures trace to mismanagement or failed control, about a quarter to governance problems of all kinds, and 53 percent to external events such as the Reformation, secularization, war, and plague. The authors attribute the stability to internal control mechanisms — value system, selection, participation, monitoring — in combination with external control, above all planned visitations by members of the order from outside the house. [limit] The paper is about agency problems in governance, not about AI or review capacity. What Chapter 14 takes from it is one arrangement: a review body placed outside the unit it reviews, arriving on a schedule.

Chapter 15 · Where Ownership Meets Capacity

  • Legal sourceThe claim that current AI rulebooks govern decision rights and none of them names a number for inflow or review time: checked against two full texts. NIST AI RMF 1.0 contains no reference to workload, throughput, time pressure, case volume, fatigue, or review capacity. The European AI regulation’s Article 14 requires understanding, monitoring, and the ability to override, and measures “proportionate to the risks, level of autonomy and context of use” — naming no quantity, no frequency, and no time per case. [limit] A negative finding from two checked frameworks, not a claim of priority. Other frameworks were not checked in full text, and the text is limited accordingly.
  • Legal sourceThe one instrument that names time and workload: Global Privacy Assembly, 47th closed session, Seoul, September 2025, Resolution on Meaningful Human Oversight of Decisions Involving AI Systems. The Resources section, verbatim: “The organization should provide the overseer with the resources necessary to adequately oversee a decision. This should include sufficient time to undertake the oversight and a reasonable workload […].” [limit] A resolution, not law. It binds nobody, and it names no number either.

Chapter 16 · The Price of the Objection

  • PreprintGreen, Ben. “The Flaws of Policies Requiring Human Oversight of Government Algorithms.” Computer Law & Security Review 45 (2022), article 105681. Examines 41 policies requiring human oversight of government algorithms. Two findings: people do not deliver the oversight demanded of them, and the requirements thereby legitimize faulty systems. Green’s own proposal is a shift from individual to institutional oversight. Freely available as arXiv:2109.05067.
  • See aboveAmazon recruiting tool — see Chapter 9.
  • Primary sourceMeehl, Paul E. Clinical versus Statistical Prediction: A Theoretical Analysis and a Review of the Evidence. University of Minnesota Press, 1954. Grove, William M., David H. Zald, Boyd S. Lebow, Beth E. Snitz, and Chad Nelson. “Clinical versus Mechanical Prediction: A Meta-Analysis.” Psychological Assessment 12, no. 1 (2000): 19–30. Kahneman, Daniel, Olivier Sibony, and Cass R. Sunstein. Noise: A Flaw in Human Judgment. Little, Brown Spark, 2021. The most fundamental challenge to treating human judgment as the benchmark; the text takes it on in the section on system trust.
  • Legal sourceThe safeguards around external audit: Sarbanes-Oxley Act of 2002, Section 301 (the audit committee is directly responsible for appointing, compensating, and overseeing the auditor) and Section 201 (prohibited non-audit services, resting on the test that a service is barred where it would have the auditor auditing its own work or performing management functions); and the SEC auditor independence rules on partner rotation, under which the lead and concurring partners may not serve the same audit client beyond five consecutive years. The European counterpart the chapter names — rotation of the audit firm after ten years, extendable to twenty after a public bidding process, with key audit partners limited to seven — is Article 17 of Regulation (EU) No 537/2014. [limit] These provisions govern the statutory audit of public companies and say nothing about reviewing AI output. This chapter transfers an arrangement; it asserts no legal duty, and the text says so.

Part Five · The Toolkit

  • The book's ownThe three doors, the minimal start, the process, and the seven worksheets are this book’s own framework.
  • The book's ownOn the difference between introducing and mandating: Haynes 2009 against Urbach 2014 — see Chapter 11. Neither study had a control group, so the contrast is a strong hint rather than proof. It is still the reason the worksheets are deliberately spare.

The Same Two Questions, Five Legal Systems

  • Editorial noteEvery position in that section was checked on September 13, 2026, and the section says so. Changes after that date are recorded, with dates, at organization-under-ai.com/cheap-to-make-costly-to-check/law.
  • Legal sourceUnited States. Equal Credit Opportunity Act and Regulation B (12 CFR 1002.9); CFPB Circulars 2022-03 and 2023-03, withdrawn May 12, 2025, as above under Chapter 4. Federal Reserve SR 11-7, replaced by SR 26-2 on April 17, 2026, as under Chapter 9. Colorado: SB 24-205 (2024); SB 25B-004, signed August 28, 2025, postponing the start to June 30, 2026; x.AI LLC v. Weiser, No. 1:26-cv-01515 (D. Colo.), order of April 27, 2026 granting the joint motion to stay enforcement, the Justice Department having intervened; SB 26-189, signed May 14, 2026, removing the duty of care, the risk management program, and the deployer impact assessments and setting a start date of January 1, 2027. Executive Order 14365, “Ensuring a National Policy Framework for Artificial Intelligence,” December 11, 2025, establishing a Justice Department AI litigation task force and making states with onerous AI laws ineligible for non-deployment funds under the BEAD broadband program.
  • Legal sourceEuropean Union. Article 22 GDPR and Case C-634/21 (SCHUFA), as under Chapter 4. Articles 14 and 26(2) of the AI Act; Directive (EU) 2024/2831 on platform work, Article 10(2). The deferral: the amending regulation of July 2026, published July 24 and in force July 27, 2026, moving the obligations for stand-alone high-risk systems to December 2, 2027 and for AI in regulated products to August 2, 2028. Audit firm rotation: Regulation (EU) No 537/2014, Article 17 — ten years, extendable to twenty after a public bidding process, with key audit partners limited to seven.
  • Legal sourceUnited Kingdom. Data (Use and Access) Act 2025, section 80, replacing Article 22 of the UK GDPR with Articles 22A to 22D; in force February 5, 2026. Solely automated significant decisions become permissible on legal bases beyond consent, subject to notification, representations, and a right to human intervention; the operative test is meaningful human involvement. Work-related stress: the Health and Safety Executive’s management standards are guidance; the duty to assess risk under which the regulator places stress is not.
  • Legal sourceGermany. § 87 (1) no. 6 BetrVG for co-determination on systems capable of monitoring behavior or performance; § 90 (1) no. 3 and § 95 (2a) BetrVG for information during planning and consent to selection guidelines, with AI named expressly since the amendment of June 18, 2021. § 5 ArbSchG for the assessment of working conditions including psychological load, without a size threshold.
  • Legal sourceJapan. Act on the Promotion of Research, Development and Utilization of AI-Related Technologies, passed May 28, 2025 — promotion-oriented, without penalties, leaving binding force to the general laws. The Act on the Protection of Personal Information contains no counterpart to Article 22 GDPR. Industrial Safety and Health Act: annual stress check for workplaces of fifty or more, in force since December 2015; amendment promulgated May 14, 2025 extending it to workplaces of any size, with a commencement set by Cabinet Order for April 1, 2028.
  • Editorial note[limit] This is the one part of the book built on a survey rather than on sources read in full. The American, European, and German instruments named here were checked against their own texts; the British, Japanese, and state-level American positions were checked against the enacting instrument and professional commentary, not against every implementing measure. The section is written coarsely for that reason, and it draws no conclusion that depends on a fine reading of any single one of them. Anyone acting on a specific duty should ask a lawyer admitted where the duty applies.

A note on what is not here

  • Editorial noteTwo kinds of entry are deliberately absent.
  • Editorial noteSources that carry no weight in the text. A number of works were reviewed and not used. Listing them would suggest a foundation the argument does not rest on.
  • Editorial noteThe state of the law beyond publication day. Everything dated sits in its own section and on the companion site, where changes since press are recorded with dates.
  • Editorial noteThe complete list, ordered by chapter and with the date each entry was last checked, is at organization-under-ai.com/cheap-to-make-costly-to-check/sources.