Thank you, Chair Gonzalez, and members of the committee, for the opportunity to testify today.
My name is Maren Hurley-Matz, and I represent New Economy Project, an economic justice organization based in New York City.
For thirty years, New Economy Project has worked with community groups to build an economy that works for all, based on cooperation, racial justice, neighborhood equity, and ecological sustainability. We have led statewide efforts to keep payday lending out of New York, end the discriminatory use of credit checks in employment, and address root causes of poverty and inequality. At the same time, we advance campaigns for community land trusts, public banks, worker cooperatives, and other bold initiatives that build community wealth. We also provide direct legal assistance to thousands of low-income New Yorkers through our free legal hotline—challenging unfair, abusive, or deceptive business practices by banks, debt collectors, fintech corporations, debt relief and credit repair companies, and other entities.
Our testimony today focuses on the New York Artificial Intelligence Act (S1169A) which would create a long-overdue regulatory framework for use of artificial intelligence systems by banking, insurance, human resources, and other industries making consequential decisions that affect people’s economic and civil rights.
AI decision-making systems are already impacting nearly every facet of New Yorkers’ lives, from surveillance pricing at the grocery store to banking and insurance underwriting. Yet these systems are being deployed with virtually no transparency, oversight, or accountability. Without strong regulation and enforcement, AI threatens to supercharge discrimination in consumer lending, debt collection, insurance, and hiring—magnifying racial inequities while making it even harder for New Yorkers to vindicate their rights.
In most contexts, New Yorkers currently have no way of knowing whether a company is using an AI decision-making system or how that system relies on their personal data. Even when consumers are aware that AI is involved, existing law generally does not require companies to disclose how AI is used, provide meaningful explanations for AI-driven decisions, or offer a clear process to verify or challenge those decisions. This lack of transparency and accountability undermines the enforcement of existing consumer protection and civil rights laws and makes unlawful or abusive business practices far more difficult to detect and remedy.
Our testimony today raises three overarching concerns about the deployment of AI in financial services.
- Unregulated AI is being used in ways that prey on low-wage workers and expand unfair and abusive lending and debt collection practices.
Unregulated AI systems are already being deployed by powerful financial and technology
companies in ways that disproportionately harm low-wage workers and communities of color.
New York has long prohibited payday lending. In recent years, however, a host of financial technology companies have deliberately designed and deployed app-based payday loan products to evade New York’s usury laws, charging exorbitant, often hidden, fees that result in triple-digit interest rates for low-wage workers. Adding insult to injury, these companies use opaque AI systems that deepen the financial harm faced by workers already living paycheck to paycheck.1
Here is how these systems work: Payday loan apps deceive people by peddling supposed “no interest” cash advances while requiring users to grant the app broad access to their bank accounts and financial data.2 This information is then fed into the app’s proprietary AI underwriting process to determine if the user will be offered a loan and, if so, at what amount.3 The same financial data is then used to monitor workers’ finances, predict when their next paycheck will arrive, and automatically debit their accounts.4 These predictions are often inaccurate, leading to premature debits that result in overdraft fees and cascading financial harm to the account holder.5
In effect, AI functions as the tool that automates, scales, and intensifies a business model designed to extract money from workers’ paychecks. New Economy Project’s research underscores the devastating impact of this predatory model, with companies extracting more than $500 million in fees from New Yorkers’ paychecks in just the past several years.6
A similar pattern is now emerging in the debt collection industry, with companies already outlining plans to use AI to more aggressively pursue debts.7 Through our NYC Financial Justice Hotline, a free legal assistance hotline, we regularly help low-income New Yorkers facing unfair and abusive debt collection practices. We see firsthand how inadequate oversight and enforcement in the debt collection industry lead to violations of due process, improperly frozen bank accounts and garnished wages, and damaged credit for low-income people, disproportionately in Black and brown communities.
Exploitative debt collectors already mislead and harass people to coerce payment regardless of their rights under state and federal law. New AI-driven debt collection tools threaten to exacerbate these harms by enabling collectors to indiscriminately scrape publicly available information, predict who is most susceptible to pressure, and target those individuals more aggressively and more frequently. Without clear liability for discriminatory AI-generated decisions, debt collectors are poised to exponentially increase surveillance and harassment of New Yorkers.
We also routinely hear from New Yorkers about inaccurate or illegitimate debts appearing on their credit reports, which can prevent people from securing housing, transportation, utilities, or insurance. AI systems are known to produce “hallucinations”—confident but false answers—to a user’s prompt. AI hallucinations in credit reporting represent an unacceptable risk to New Yorkers’ access to fair credit.
This problem is not limited to credit reporting; the proliferation of AI chatbots has increased the risk that people will receive incorrect or misleading financial or legal information. One recent hotline caller shared that an AI system miscalculated the period between 2021 and 2023 as three years, instead of two, leading her to believe her debt was time-barred under NY’s three-year statute of limitations when it was not.
Understandably, many New Yorkers may reasonably rely on AI-generated legal information and be misled about their rights. The AI chatbot bill (S5668), also before this committee, would establish accountability for harmful information produced by AI chat systems. Together, these measures are necessary to ensure that AI does not undermine New Yorkers’ rights or further entrench racial and economic inequality in our state.
- New York should unequivocally reject biased “black box” AI decision-making as
the status quo.
It is well established that any AI system is only as good as the data it is trained on—garbage in, garbage out. Due to the enduring legacies of bank redlining and segregation, financial services data is inherently biased. AI systems trained on historical financial data continue to reproduce and deepen racial disparities in banking and financial services.
A 2022 UC Berkeley study found that fintech consumer lending AI charged communities of color an extra $450 million in interest each year, compared to white borrowers.8 Multiple lawsuits have been brought against State Farm after a study by NYU’s Center on Race, Inequality, and the Law revealed that the company’s AI fraud detection systems discriminated against claims filed by Black policyholders, resulting in Black homeowners paying higher costs for inferior insurance coverage.9 Myriad other examples demonstrate the pattern of discriminatory and abusive decision-making by AI. New Yorkers must have the right to opt out of AI decision-making for all banking and financial services.
Where AI is offered as an option, individuals should not have to choose between fairness and speed, particularly in time-sensitive situations like buying a car or obtaining a loan. It is not enough for companies to disclose that AI will be used to analyze their personal data without clearly identifying the specific information that the companies’ AI will use to make a consequential decision.
To prevent discrimination, New Yorkers must be able to understand and verify the basis for a company’s decision. This principle is already recognized in the Equal Credit Opportunity Act, which requires creditors to provide statements of specific reasons to applicants against whom adverse action is taken. Such transparency is even more vital in the context of AI, which uses data points such as location, device manufacturer, text formatting habits, or purchase patterns—that can serve as proxies for protected characteristics like race, gender, and family status.10
The Department of Financial Services has already recognized these risks in the insurance sector. In a 2024 circular letter, DFS required insurers using AI to demonstrate that proxy factors were not used in underwriting or pricing, mandated disclosure to consumers when AI is used and what data sources are involved, and affirmed consumers’ right to request specific information about adverse decisions.11 Developers and deployers in any industry must not be permitted to rely on opaque, “black box” AI models to make decisions impacting New Yorkers’ fundamental rights or access to essential services.
- Trump’s attack on states’ AI regulation underscores the need for notice,
transparency, and a private right of action.
Just over a month ago, President Trump signed an Executive Order attacking state laws regulating AI. This dramatic overreach of executive power, following the assault on the Consumer Financial Protection Bureau, is yet another clear signal that the federal government will not act to protect consumers from exploitation or discrimination by Big Tech. This represents an abrupt reversal from the common-sense AI protections adopted during the Biden administration, when the CFPB, DOJ, EEOC, and other federal agencies issued guidance regarding companies’ use of AI and vigorously investigated discrimination.12
States are now the primary bulwark preventing a race to the bottom on AI. Although New York recently enacted the RAISE Act, that law does not meaningfully address AI decision-making systems already harming New Yorkers.
In this moment, it is critical that New Yorkers have notice, transparency, and the ability to hold irresponsible actors accountable through a private right of action. Only with these tools can New Yorkers meaningfully protect themselves from discriminatory and abusive uses of AI. The NY AI Act establishes these baseline protections by giving the right to opt out of AI decision-making, requiring companies to provide information necessary to understand and appeal adverse AI-driven decisions, and creating a private right of action so New Yorkers can hold discriminatory actors accountable. The bill’s audit and reporting requirements are essential to ensure AI systems are tested for bias and reliability, and its prohibition on AI social scoring systems addresses some of the most dangerous uses of automated decision-making.
At the same time, the bill should go further. The notice provisions should require mandatory disclosures informing users of the risks of AI use, companies’ liability for discriminatory outcomes, or individuals’ rights under the law. In addition, the 45- to 90-day investigation period for appeals unfairly burdens people facing time-sensitive harms and should be shortened. We look forward to working with the Committee to address these concerns and strengthen the New York AI Act.
Through robust consumer protections and vigorous state-level enforcement, New York has long been a national leader. The state must continue to lead on AI regulation and ensure that no New Yorker faces discrimination by artificial intelligence in financial services or other consequential economic decisions that affect people’s lives.
Thank you again for the opportunity to testify today.
1 New York v. MoneyLion, CV-451303-25 (April 14, 2025); Cleo AI Inc.,
https://web.meetcleo.com/cash-advance, accessed 1/13/2026 (“With our AI cash advance app, we know that when you’re in need of funds, you can’t afford to waste time. Access up to a $250 cash advance.”)
2 New York v. DailyPay, CV-154851-25 (April 14, 2025), paragraph 40; New York v. MoneyLion,
CV-451303-25 (April 14, 2025), paragraphs 40-46.
3 Id.
4 New York v. MoneyLion, CV-451303-25 (April 14, 2025), paragraphs 44-46.
5 New York v. MoneyLion, CV-451303-25 (April 14, 2025), paragraphs 44-46; Lucia Constantine, et al., “Not Free: The Large Hidden Costs of Small-Dollar Loans Made Through Cash Advance Apps,” Center for Responsible Lending (April 2024) https://www.responsiblelending.org/sites/default/files/nodes/files/research-publication/crl-not-free hidden-costs-apr2024.pdf.
6 New Economy Project, “$500 Million and Counting: How Unchecked Payday Loan Apps Are Draining New Yorkers’ Paychecks,” (March 2025)
https://www.neweconomynyc.org/wp-content/uploads/2025/03/500-Million-and-Counting.pdf
7 Ohad Samet, “AI is set to permanently disrupt the debt collection industry,” American Banker (Oct. 10, 2025).
8 Robert Bartlett, et al., “Consumer-lending discrimination in the FinTech Era,” Journal of Financial
Economics (Jan. 2022).
9 Chad Hemenway, “Lawsuit Against State Farm Alleges Racial Discrimination in Claims Handling,”
Insurance Journal (Dec. 16, 2022).
10 American Academy of Actuaries, “Discrimination: Considerations for Machine Learning, AI Models, and Underlying Data,” (Feb. 2024)
https://www.actuary.org/sites/default/files/2023-08/risk-brief-discrimination.pdf.
11 NY Department of Financial Services, “Insurance Circular Letter No. 7 regarding the Use of Artificial Intelligence Systems and External Consumer Data and Information Sources in Insurance Underwriting and Pricing,” (July 11, 2024) https://www.dfs.ny.gov/industry-guidance/circular-letters/cl2024-07
12 Jory Heckman, “EEOC, DOJ ‘sounding alarm’ over AI hiring tools that screen out disabled applicants”, May 12, 2022; Consumer Financial Protection Bureau, “CFPB Issues Guidance on Credit Denials by Lenders Using Artificial Intelligence,” (Sept. 19, 2023); Equal Employment Opportunity Commission, “iTutorGroup to Pay $365,000 to Settle EEOC Discriminatory Hiring Suit,” (Sept. 11, 2023).