The Algorithm Is Your New Boss

The Resume Black Hole. You are a highly qualified Black woman who has applied for several senior marketing roles at a large tech company but have never once received an interview, while your less-experienced white male colleagues get calls back. You suspect the company's AI resume screener, trained on a history of biased hiring, is systematically downgrading your resume. This is a classic case of potential disparate impact under Title VII. The Video Interview Rejection. You have a mild speech impediment that has never affected your job performance. After a recorded video interview where you felt you answered every question perfectly, you receive an automated rejection. The company uses an AI to score candidates on "communication skills," and you believe it penalized your speech patterns. This is a potential violation of the ADA. The Productivity Firing. You work at a warehouse where your every move is tracked. You are a consistent performer but had to take more frequent bathroom breaks one week due to a temporary medical issue. You are fired at the end of the week for "failing to meet productivity targets," a decision recommended by the company's management algorithm. This could be a case of disability discrimination and a failure to provide reasonable accommodation. The Phantom Job Ad. You are a 58-year-old woman actively looking for work on social media. You notice that your 30-year-old daughter sees numerous job ads for a local tech startup in her feed, but you see none, even when you search for the company directly. The company may be using the platform's ad-targeting tools to exclude older workers from their recruiting pipeline, a potential violation of the ADEA. The Secretive System. You are denied a promotion based in part on a "low engagement score" generated by the company's bossware, which analyzes your email response times and chat activity. You ask HR how the score is calculated, and they tell you it is a "proprietary vendor algorithm." Depending on your location (e.g., New York City), you may have a legal right to more information about how that automated decision was made. The Black Box Problem One of the greatest challenges in fighting algorithmic discrimination is the "black box" nature of many advanced AI systems. Deep learning and neural network models can process millions of data points to reach a decision, but it is often impossible , even for the developers themselves , to fully explain how a specific output was derived from a given input. When an AI tool rejects your job application or flags you for termination, the system may be incapable of providing a human-understandable reason. This opacity creates three critical legal problems. First, employers cannot defend their decisions if they cannot explain them. When an employee asks "why was I fired?" and the answer is "the algorithm determined you were not a good fit," that is not a legally sufficient justification. Second, the black box makes meaningful bias audits extraordinarily difficult, because auditors cannot examine the internal logic that produces discriminatory outcomes. Third, algorithmic opacity fundamentally undermines due process , workers have a right to understand the basis for adverse employment actions, and "the computer said so" does not satisfy that right. Courts are beginning to grapple with this problem. In AI discrimination litigation, discovery battles over algorithmic transparency are becoming the central front. Plaintiffs demand access to the algorithm's source code, training data, and decision logic. Defendants resist, claiming trade secret protection. The resolution of these battles will determine whether workers can meaningfully challenge algorithmic discrimination or whether the black box becomes an impenetrable shield for employer liability. Disparate Treatment in the Age of AI While disparate impact , the unintentional discriminatory effect of a facially neutral tool , receives most of the attention in AI employment law, disparate treatment remains a critical theory of liability. Disparate treatment occurs when an employer intentionally uses a protected characteristic in its decision-making. In the AI context, this can happen in two ways. The first is explicit programming. If an employer or vendor programs an AI tool to consider age, race, or gender as a factor in employment decisions, that is intentional discrimination, regardless of the technological sophistication of the tool. The EEOC v. iTutorGroup case was essentially a disparate treatment case: the company's AI was programmed with age cutoffs that automatically rejected women over 55 and men over 60. The second, more insidious form occurs when an employer knows that an AI tool is producing discriminatory results but continues to use it anyway. If an employer receives a bias audit showing that its AI hiring tool selects Black applicants at only 60% the rate of white applicants, and the employer continues using the tool without modification, a court could find that the employer's continued use constitutes intentional discrimination. Knowledge plus inaction equals intent. This is why the transparency and audit requirements in the Colorado AI Act and New York City's Local Law 144 are so important , they create a documented record of what the employer knew and when they knew it. The Multi-Agency Enforcement Landscape The fight against algorithmic discrimination is not limited to the EEOC. Multiple federal and state agencies are now scrutinizing AI in the workplace, creating a multi-front enforcement landscape that workers should understand and leverage. The Federal Trade Commission entered the arena after the ACLU filed a complaint alleging that AON's AI-powered assessment tools discriminated against job applicants. The FTC's investigation signals that consumer protection laws may provide an additional avenue for challenging AI-related employment discrimination, particularly when AI vendors market their products with claims about fairness and objectivity that prove false. The National Labor Relations Board has identified AI-powered workplace surveillance as a threat to workers' Section 7 rights. When employers deploy AI tools that monitor communications, track organizing activity, and analyze worker sentiment, the NLRB views this as a potential unfair labor practice that chills protected concerted activity. This is particularly relevant for workers who are organizing, filing complaints, or discussing working conditions. State agencies are also stepping up. California's Civil Rights Department has approved regulations specifically addressing AI-related employment discrimination. Illinois's Department of Human Rights enforces the state's expanded AI anti-discrimination framework under HB 3773. The Texas Workforce Commission oversees compliance with the Responsible AI Governance Act. Workers should not limit themselves to a single agency , filing with multiple agencies simultaneously can maximize pressure and increase the likelihood of a meaningful investigation. References [1] SoftwareSeni. (2026, January 15). Understanding Employee Monitoring Software and the Rise of Workplace Bossware in 2026. https://www.softwareseni.com/understanding-employee-monitoring-software-and-the-rise-of-workplace-bossware-in-2026/ [2] Reuters. (2018, October 10). Amazon scraps secret AI recruiting tool that showed bias against women. https://www.reuters.com/article/world/insight-amazon-scraps-secret-ai-recruiting-tool-that-showed-bias-against-women-idUSKCN1MK0AG/ [3] U.S. Equal Employment Opportunity Commission. (2023, April 25). EEOC, DOJ, CFPB, and FTC Joint Statement on AI. https://www.eeoc.gov/newsroom/eeoc-doj-cfpb-and-ftc-joint-statement-ai [4] Brookings Institution. (2025, April 25). Gender, race, and intersectional bias in AI resume screening. https://www.brookings.edu/articles/gender-race-and-intersectional-bias-in-ai-resume-screening-via-language-model-retrieval/ [5] Stanford University. (2025, October 17). Researchers uncover AI bias against older working women. https://news.stanford.edu/ stories/2025/10/ai-llms-age-bias-older-working-women-research/ [6] HR Dive. (2025, March 24). AI hiring software was biased against deaf employees, ACLU alleges. https://www.hrdive.com/news/ ai-intuit-hirevue-deaf-indigenous-employee-discrimination-aclu/ 743273/ [7] ExpressVPN. (2024). 2024 Survey: AI-Powered Workplace Surveillance. https://www.expressvpn.com/ (Note: Specific URL for the survey is illustrative). [8] The Guardian. (2024, May 21). 'You feel like you're in prison': workers claim Amazon's surveillance is used to deter union efforts. https://www.theguardian.com/us-news/article/2024/may/21/ amazon-surveillance-lawsuit-union [9] K&L Gates. (2023, May 31). EEOC Issues Nonbinding Guidance on Permissible Employer Use of Artificial Intelligence. https://www.klgates.com/EEOC-Issues-Nonbinding-Guidance-on-Permissible-Employer-Use-of-Artificial-Intelligence-to-Avoid-Adverse-Impact-Liability-Under-Title-VII-5-31-2023 [10] U.S. Equal Employment Opportunity Commission. (2023, September 11). iTutorGroup to Pay $365,000 to Settle EEOC Discriminatory Hiring Suit. https://www.eeoc.gov/newsroom/itutorgroup-pay-365000-settle-eeoc-discriminatory-hiring-suit [11] U.S. Equal Employment Opportunity Commission. (2022, May 12). The Americans with Disabilities Act and the Use of Software, Algorithms, and Artificial Intelligence. https://www.eeoc.gov/ laws/guidance/americans-disabilities-act-and-use-software-algorithms-and-artificial-intelligence [12] U.S. Equal Employment Opportunity Commission. (2024, April 9). Brief in Mobley v. Workday, Inc. https://www.eeoc.gov/ litigation/briefs/mobley-v-workday-inc The Road Ahead The fight against algorithmic injustice is just beginning. In December 2025, Executive Order 14365 signaled a federal preference for deregulating AI, explicitly aiming to preempt the patchwork of state laws that were just beginning to hold employers accountable for their automated systems. But this executive order does not rewrite the civil rights laws that have protected American workers for over half a century. Title VII, the ADA, the ADEA, and the FLSA remain in full force. There is no AI exemption. In early 2024, Swedish fintech company Klarna announced that its AI assistant , built in partnership with OpenAI , was handling the work of 700 customer service agents. The company's CEO presented this as a triumph of efficiency, a proof of concept that AI could replace human workers at scale. Seven hundred people lost their jobs. The algorithm had clocked in. But the story did not end there. By mid-2025, Klarna was struggling. The AI's performance, it turned out, was not as strong as the press releases had claimed. Customer satisfaction had declined. The nuance, empathy, and judgment that human agents brought to complex customer interactions could not be replicated by a chatbot. Klarna needed human workers again. But instead of rehiring them as employees with benefits and job security, the company reportedly explored an "on-demand" model , bringing workers back as independent contractors, available when needed, discarded when not. This is a pattern every worker should understand, because it reveals the full cycle of algorithmic displacement. First, the company replaces you with AI. Second, the AI underperforms. Third, the company rehires humans , but on worse terms, with fewer protections, and at lower cost. The worker who was once an employee with benefits becomes a contractor with none. And the legal protections that applied to the employee , overtime, anti-discrimination, the right to organize , may not apply to the contractor at all. If this happens to you, the legal question is whether the reclassification is legitimate or whether you are being misclassified. As we discussed in Chapter 9, the distinction between an employee and an independent contractor is not a label the company gets to choose. It is a legal determination based on the economic reality of the relationship. If the company controls how you do your work, when you do it, and what tools you use , the same way they did when you were an employee , then you may still be an employee under the law, regardless of what they call you. Document the similarities between your old role and your new one. That documentation could be the foundation of a misclassification claim. The displacement of workers by AI-assisted decision-making is not limited to the private sector. In 2025, the federal government's Department of Government Efficiency initiative used AI-assisted processes to identify federal positions for elimination, contributing to the departure of more than 211,000 federal employees in a single year. Whatever one's views on the size of the federal workforce, the scale of this action demonstrates a critical point: when AI is used to make workforce reduction decisions at speed and scale, individual workers can be swept away before they have a chance to understand what happened, let alone fight back. The tools in this book , documentation, legal deadlines, protected class analysis , apply whether your employer is a private company or the federal government. [Footnote:] In October 2024, the European Union adopted the Platform Work Directive, which requires gig economy platforms to be transparent about how their algorithmic management systems work, gives workers the right to challenge automated decisions, and mandates human review of significant employment decisions. While not binding in the United States, the Directive represents a model that American advocates and lawmakers are actively studying. Its core principle , that a worker managed by an algorithm deserves the same due process as a worker managed by a human , is one that American law will eventually need to confront. At the state level, the momentum is accelerating. Illinois has expanded its AI Employment Act to cover all automated decision-making tools. Colorado's AI Act takes effect in June 2026. Texas has enacted its own Responsible AI Governance Act. New York City's Local Law 144 continues to require bias audits for automated hiring tools. California is advancing sweeping AI transparency legislation. These laws are not going away, and they are getting stronger. The courts are opening new fronts as well. Landmark decisions have established that AI vendors themselves can be held liable as agents of the employer (see Chapter 21). The legal front against AI hiring tools is expanding beyond anti-discrimination law into entirely new territory. In January 2026, a class-action lawsuit was filed against Eightfold AI, Inc., one of the most prominent AI-powered hiring platforms in the country. The allegation is striking: the lawsuit claims that Eightfold scrapes data on over one billion workers worldwide to create secretive candidate "scores" , numerical ratings that determine whether a job applicant advances or is silently rejected , without the workers' knowledge or consent. The legal theory is novel and powerful. The plaintiffs argue that these candidate scores function as consumer reports, and that Eightfold is operating as a consumer reporting agency under the Fair Credit Reporting Act (FCRA). If this theory holds, it means that AI hiring platforms must comply with the FCRA's requirements: transparency about what data is being collected, accuracy in the information used, and the right of individuals to dispute and correct errors. This is a different legal weapon than Title VII or the ADA. It does not require you to prove discrimination. It requires only that the company collected and used your data without following the rules. This case matters because it gives workers a new avenue. If you have been rejected for jobs through an AI hiring platform and suspect that a hidden score or profile is following you from application to application, the FCRA may give you the right to see that score, challenge its accuracy, and hold the platform accountable. Watch this case closely. Its outcome could reshape the entire AI hiring industry. The algorithm is not just deciding who gets hired and who gets fired. It is deciding what you get paid , and it may be paying you less than the person doing the same job, based on nothing more than its prediction of your desperation. A 2025 Human Rights Watch report documented a practice that researchers call algorithmic wage discrimination on gig economy platforms. The report found that algorithms dynamically adjust pay rates for individual workers based on their behavioral patterns. If you have recently accepted several jobs in a row, the algorithm infers that you are willing to work for less , and it offers you a lower rate on the next job. If you have been idle, declining offers, the algorithm infers that you need a higher incentive , and it offers you more for the same work. Two drivers, picking up the same order from the same restaurant and delivering it to the same neighborhood, may be paid different amounts based on nothing more than the algorithm's assessment of how much each one needs the money. This is not traditional wage theft. There is no employer shorting your overtime or skimming your tips. This is something newer and, in many ways, more insidious: a system that uses your own behavior as a weapon against your earnings. The law has not yet fully caught up to this practice. But the documentation is the first step. If you work on a gig platform, track your pay rates over time. Screenshot your offers. Compare notes with other workers doing the same jobs. The patterns you find may become the evidence that drives the next wave of litigation. The EEOC continues to issue guidance making clear that employers cannot hide behind algorithmic opacity. And a growing wave of litigation, from the iTutorGroup settlement to the Sirius XM complaint, is building a body of case law that will define the rights of workers in the algorithmic age. This chapter has given you the knowledge to understand how the algorithm works, where the bias comes from, and what the law says about it. The rest of this book has given you the tools to fight back. The algorithm may be your new boss, but neither it nor the company that deployed it is above the law. CLOCKED IN Clocked In AI is not exempt from employment law. Title VII, the ADA, the ADEA, and the FLSA all apply with full force to automated systems, and your employer is liable for discrimination caused by their AI tools, even if a third-party vendor built them. The Four-Fifths Rule is your statistical weapon. If an AI system selects members of a protected class at a rate less than 80% of the rate for the most-selected group, it creates a presumption of illegal disparate impact that the employer must rebut. Document everything the algorithm does to you. Keep a detailed log of every interaction with an automated system, every productivity score, every scheduling decision, and every automated rejection. This digital paper trail is the evidence you will need to challenge the algorithm. Before You Move On ▢ ▢ ▢ ▢ ▢ CH A P T E R 2 1 A C T I O N C H E C K L I S T I understand that AI tools are subject to the same anti-discrimination laws as human managers If I suspect AI was used in a decision about me, I have documented what I know about the system I have checked whether my state or city has AI-specific employment laws I know about the Mobley v. Workday and Eightfold AI cases and what they mean for my rights I understand that I can file an EEOC charge even if an algorithm made the decision SECTION SUMMARY This chapter examines the growing role of artificial intelligence (AI) in the workplace, from hiring and promotions to terminations. It explains how seemingly objective algorithms can perpetuate and even amplify existing biases, leading to unlawful discrimination. You will learn about your legal rights under federal, state, and local laws when decisions are made by AI, including how to identify disparate impact using concepts like the "Four-Fifths Rule." Ultimately, this section provides the knowledge to challenge unfair automated decisions and hold employers accountable.