The New York City Council passed Intro. 1174-A on March 20, 2025, banning the use of personal data for “surveillance pricing” — a law that directly implicates AI-driven dynamic pricing and personalization systems. The bill, flagged in a Hacker News discussion, targets algorithmic price discrimination based on a customer's data footprint.
Law: Intro. 1174-A | Vote: 44-5 | Effective: Potentially mid-2026
Scope: NYC businesses | Penalty: Up to $500 per violation
What the Law Prohibits
The law makes it illegal for sellers in New York City to use, share, or sell personal identifying information to charge different prices for goods or services. This specifically outlaws practices where algorithms adjust prices based on data like browsing history, location, purchase history, or device type.
Protected information includes name, address, email, financial data, and unique device identifiers. The law carves out exceptions for discounts based on membership programs (like Costco), senior/student discounts, volume-based pricing, and prices negotiated through bids.
How It Works and Enforcement
Enforcement is designed to be consumer-driven. The law creates a private right of action, allowing individuals to sue for damages of $500 per violation, plus attorney’s fees. This follows a similar model to NYC’s biometric privacy law, which has led to significant litigation.
The city’s Department of Consumer and Worker Protection can also promulgate rules and enforce the law. Fines collected will be directed to the city’s general fund.
Bottom line: A plaintiff-friendly law that puts the onus on businesses to prove their pricing algorithms don't use protected personal data.
"Key Legal Definitions"
The law defines "Surveillance Pricing" as charging different prices based on a person's "personal identifying information." "Personal Identifying Information" is broadly defined to include:
The Numbers: Council Support and Timeline
The bill passed with overwhelming support in the 51-seat council, with a final vote of 44 in favor, 5 against, and 2 abstentions. It now moves to Mayor Eric Adams’ desk. If signed, the law would take effect 180 days after signing, placing potential enforcement in late 2025 or early 2026.
The Hacker News thread garnered 43 points and 9 comments, with discussion focusing on the law's technical feasibility and its implications for common e-commerce practices like coupon targeting.
Pros and Cons for AI Practitioners
Potential Benefits:
- Clearer Boundaries: Provides a legal bright line against the most invasive forms of algorithmic price discrimination.
- Consumer Trust: May increase user trust in platforms that visibly comply, mitigating "algorithmic resentment."
- Uniformity: Creates a single compliance standard for businesses operating in NYC, rather than a patchwork of internal policies.
Significant Drawbacks:
- Compliance Complexity: Distinguishing between prohibited "personal identifying information" and permitted "aggregated data" for model training is a technical and legal challenge.
- Litigation Risk: The $500 per violation private right of action creates a high-risk environment for class-action lawsuits.
- Blunt Instrument: May hinder beneficial personalization, like offering discounts to price-sensitive cohorts identified by ML models.
Alternatives and Regulatory Comparisons
NYC’s approach is more narrowly targeted than broader algorithmic accountability laws but is highly enforceable. Other frameworks take different tacks.
| Feature | NYC Intro. 1174-A (Surveillance Pricing) | Colorado AI Act (Broad Accountability) | EU AI Act (Risk-Based) |
|---|---|---|---|
| Scope | Pricing using personal data | All "high-risk" AI systems | All AI systems by risk level |
| Enforcement | Private lawsuits + city fines | Attorney General enforcement | National authorities, large fines |
| Developer Duty | Avoid using PII for pricing | Risk management, impact assessments | Conformity assessments, transparency |
| Status | Passed Council, awaiting signature | Signed into law (Effective Feb 2026) | Fully enacted |
Existing tools like IBM’s AI Fairness 360 or Microsoft’s Fairlearn can help audit for bias but are not designed for this specific legal compliance. The closest operational alternative is differential privacy techniques, which could allow for aggregate trend analysis without using individually identifiable data.
Who Should Use This Framework (And Who Should Worry)
This law is immediately relevant for:
- E-commerce and Retail Tech Teams building dynamic pricing or personalized offer engines for NYC customers.
- SaaS Platforms that provide pricing optimization tools to NYC-based businesses.
- Legal & Compliance Teams at any company selling digitally in NYC, who must now audit pricing models.
Less impacted are:
- Developers working on non-commercial or non-pricing AI (e.g., content recommendation, creative tools).
- Companies with simple, fixed pricing or membership-based discount models that fall under the explicit exceptions.
- Businesses with no NYC customer base.
The Bottom Line for AI Development
Intro. 1174-A is a targeted strike against a specific, unpopular use of personal data. It won't stop all algorithmic pricing, but it forces a decoupling of pricing logic from individual user profiles. For developers, the immediate task is data governance: ensuring pricing models cannot access or are not trained on the protected categories of personal identifying information.
The law signals a growing legislative comfort with surgically regulating specific AI applications rather than waiting for comprehensive federal rules. Similar cities are likely to follow, making NYC's experiment a potential blueprint.
Top comments (0)