Skip to main content

How to Advertise to Newly Elected Officials Before Day One

Learn more
Policy Pulse Nov 10, 2026

New York Bill Would Make Frontier AI Developers Strictly Liable

A new New York bill would make developers of the largest AI models strictly liable for some harms to people who never used the model. Four technology bills were introduced in election week and none moved further, while Milwaukee and El Paso County took formal votes on data centers.


What Changed

  • New York · S.10733, the Understanding Artificial Intelligence Act, makes developers of “covered models” strictly liable for certain injuries to non-users · Introduced Nov. 6 by NY Sen. Rachel May (D-NY-048) and referred to Rules.
  • Nevada · SB31, filed on behalf of the Attorney General, sets rules for AI companions and bans surge and surveillance pricing at large grocery stores · Prefiled Nov. 4.
  • Pennsylvania · H.B.2888, the Data Center Community Support and Compliance Act, sets energy, water, reporting and ratepayer requirements for hyperscale data centers · Introduced Nov. 4 (referred to Energy).
  • Tennessee · SB0003, the Agricultural Data Privacy Act, requires a producer’s consent before anyone sells that producer’s agricultural data · Filed for introduction Nov. 9.

No bills on social media, age verification, app stores or kids’ online safety changed status in Quorum this week.


The Lead

New York S.10733: liability for large AI model developers. Introduced Nov. 6 by NY Sen. Rachel May (D-NY-048). The bill says it “shall take effect immediately” if enacted.

The core rule makes developers liable for harms they didn’t intend or foresee:

“Except with respect to any causes of action for defamation, developers of covered models shall be strictly liable, regardless of the degree of care they exercised, for all injuries to a non-user of the covered model that satisfy the actual harm element of an ordinary negligence claim if: (a) those injuries are factually and proximately caused by a covered model that engages in conduct that, if undertaken by an adult human of sound mind, would satisfy the elements of negligence or any intentional tort or crime; and (b) that conduct was neither intended nor could have been reasonably anticipated by: (i) the user of the model; or (ii) any intermediary that fine-tuned, scaffolded, or otherwise modified the model.”

  • Which models count. A covered model is one trained with more than 10 to the 26th power operations, at a compute cost above $100 million.
  • Mental state. Where a tort requires intent or a similar mental state, the bill creates “a rebuttable presumption that the artificial intelligence system satisfies the relevant mental state” if a person acting the same way would have had it. It adds: “It shall not be a defense that artificial intelligence systems are incapable of having mental states.”
  • Defenses. A developer can show the model met “the standard of care applicable to humans who perform the same function,” or that the harm came from “a capabilities failure,” where the model fell short of what the user intended or could reasonably expect.

Nevada SB31: AI companions and grocery pricing. Prefiled Nov. 4 on behalf of the Attorney General.

  • AI companions. The bill covers systems “designed to simulate a sustained human or human-like relationship with a user.” Operators must keep protocols to “detect and address self-harm and suicidal ideation by a user” and tell users clearly that they are “communicating with an automated system and not with a human.”
  • Minors. For users under 18, operators must block outputs that encourage them to keep the interaction secret, “self-isolate or refrain from seeking help from a health professional or appropriate adult.”
  • Exemptions. Systems built for business uses, such as customer service, productivity, internal research and technical help, are exempt.
  • Grocery pricing. Stores over 10,000 square feet could not use surge pricing or surveillance pricing, meaning prices “based on personally identifiable information or inferred data,” with some exceptions. The penalty under the state’s Unfair Trade Practice Act would rise to “$50,000 for each violation.”

Pennsylvania H.B.2888: hyperscale data centers. Introduced Nov. 4 by PA Rep. Joe Webster (D-PA-150) and cosponsors.

  • Ratepayers. Before operating, an owner must certify that changes to its energy plan have “not imposed additional costs on other customers of the electric distribution company.”
  • Clean power. A shortfall in required “incremental clean firm energy” triggers a compliance payment, set each year at double the average cost of solar power.
  • Water. Annual reports must certify that the data center’s “water use has not adversely impacted the quantity or quality of water for another use or user.”

Our read: New York’s bill shifts the frontier AI debate from disclosure toward liability, with a compute and cost threshold and a presumption about an AI system’s mental state. AI companion and chatbot safety is a live 2027 track in at least three states: Nevada’s new bill, an Ohio committee’s work on AI self-harm rules and a Virginia commission’s study of chatbot companionship.


In Committee

Most committee activity was end-of-year study work. Ohio’s House technology committee was the exception. No technology bills cleared committee.

  1. Ohio House Committee on Technology and Innovation, Nov. 9. At its fifth hearing on HB524, which would penalize AI models that suggest a user harm themselves or someone else, the committee adopted an amendment without objection. The bill was not voted out. Vice Chair OH Rep. Heidi Workman (R-OH-072) said the amendment sends “50% of fines imposed under the bill on companies deploying artificial intelligence models” to the Consumer Protection Enforcement Fund and 50% to the 988 crisis line fund. Chair OH Rep. Thaddeus “Thad” Claggett (R-OH-068) said he doesn’t “always know how to, to balance the forces between the positive gains that you guys have in your bills and the downside risk.” Workman · Claggett
  2. Virginia Joint Commission on Technology and Science, Nov. 4. Members split on independent verification for AI. The presentation recommended that “Virginia begin with a limited risk-based pilot.” VA Del. Bonita Anthony (D-VA-092) moved not to endorse the pilot, citing “insufficient evidence … or fiscal capacity.” VA Sen. Barbara Favola (D-VA-040) said, “I ultimately believe a pilot is probably where we need to end up.” The motion that carried sends the question to a work group, which will report at the commission’s first meeting in 2027. Recommendation · Anthony · Favola · Motion
  3. Rhode Island commission on technology in schools and mental health, Nov. 4. A witness described the tradeoff in age checks for school apps: “figuring out who is under 13 requires more data collection. So there’s a paradox there, and that’s something our commission will have to confront.” Testimony

State And Local Spotlight

Data centers drew formal votes in Milwaukee and El Paso County, and Pennsylvania’s new bill proposed state rules for the same projects.

  1. Milwaukee, Wis., Common Council, Nov. 4. Voted 15-0 to direct the Department of City Development to recommend policies and regulations “to prevent negative environmental, neighborhood, and community impacts of data processing centers.” It also voted 15-0 to create a data center dashboard. Minutes
  2. El Paso County, Texas, Commissioners Court, Nov. 5. Voted 5-0 to ask the Army to name the county a “cooperating agency” in its environmental review of a proposed “commercial data center development on approximately 1,384 acres at Fort Bliss.” Minutes
  3. Pennsylvania House. H.B.2888 would require state certifications on energy, water and community benefits before a hyperscale data center is built or starts operating. See The lead.
  4. Puerto Rico. PC0777 would require “adequate age verification mechanisms” before minors can reach sexually explicit sites. It reached the executive stage Nov. 2, and the House president signed it Nov. 5. It isn’t counted in this week’s movement because its stage didn’t change in the window.
  5. Tampa, Fla., City Council, Nov. 5. The agenda included a staff resolution to stop accepting data center applications under Florida’s pending ordinance doctrine and to schedule hearings on a moratorium ordinance for Dec. 3 and Jan. 7. Agenda

Quorum found no formal local votes on AI policy itself this week. Santa Clara County held its generative AI policy item until Nov. 17.


What Lawmakers Are Saying

  1. Pennsylvania: AI deepfake reporting. Four House Republicans posted Nov. 3–5 about Act 87 of 2026, which “requires mandated reporters of child abuse to immediately notify law enforcement when they become aware of suspected child sexual abuse material involving artificial intelligence-generated deepfake images.” PA Rep. Joseph “Joe” D’Orsie (R-PA-047) added that the law “will take effect in late January.” D’Orsie
  2. Illinois: working across states. IL Sen. Mary Edly-Allen (D-IL-031), Nov. 5, on joining NCSL’s Task Force on Artificial Intelligence, Cybersecurity and Privacy: “This is not about stopping innovation, but balancing the great promise of AI with its potential harms.” Post
  3. Kansas: against age verification mandates. KS Rep. Rebecca Schmoe (R-KS-059), Nov. 5: “Age-verification mandates are a government-mandated family registry living indefinitely inside a commercial platform that will follow your child everywhere they go for the rest of their life.” Post

How often officials posted on these topics (Facebook and Instagram):

  • 15 posts by officials mentioned “artificial intelligence,” about half the prior week’s count.
  • 2 posts mentioned “data privacy,” down from 3.
  • 1 post mentioned “age verification,” up from 0.

Election posts made up much of this week’s volume.


What To Watch

  • Nov. 17, 2026, 9:30 a.m.: Santa Clara County, Calif., Board of Supervisors is due to take up its “Generative Artificial Intelligence Policy.” Quorum
  • Next step: Puerto Rico’s age verification bill, PC0777, now awaits the governor’s action.
  • Dec. 3, 2026, 5:01 p.m., and Jan. 7, 2027, 10 a.m.: Tampa’s scheduled public hearings on a data center moratorium ordinance.
  • Late January 2027: Pennsylvania’s Act 87 deepfake reporting rule takes effect.

No hearings are on the calendar through Dec. 10 for S.10733, SB31 or H.B.2888.


Movement

Stage Bills Jurisdictions
Signed 0 None
Passed a chamber 0 None
Cleared committee 0 None
Introduced 4 Nevada (1), New York (1), Pennsylvania (1), Tennessee (1)

Source: Quorum legislative data, Nov. 3–9, 2026. Each bill is counted once, at its furthest step this week. Track bills like these with Quorum State and Quorum Federal.