reportAI in CourtsAI Governance
41 US AI bills that reach courts, evidence and legal practice, 2024–2026
41 verified US bills and laws as of 7 Sep 2026: 10 in force, 23 pending, 8 not advanced, with court and legal-practice duties.
Edited and verified by Cognesio LLP
Researched with AI assistance · sources verified by Cognesio LLP · How this was made ↓
As of 7 September 2026, the SafeLegalAI legislation register holds 41 verified state and federal bills and laws from the United States whose operative sections reach court proceedings, court administration, evidence, legal services, attorneys, police reports or legal aid. Ten are in force, 23 are still pending or enrolled, and eight have died or had a veto sustained.
The current live question is California SB 574, which passed both houses on 31 August 2026 and was enrolled on 4 September 2026. Presentation to the Governor had not yet been recorded when the LegInfo history and status pages were read on 7 September 2026; once presented, the Governor has until 30 September 2026 under Cal. Const. art. IV §10(b)(2). It would add Business and Professions Code §6068.1(a), amend Code of Civil Procedure §128.7, add arbitrator provisions and direct the Judicial Council to revisit its generative-AI standard. The bill page is in the register; the official history and text are on California’s legislature site.
Key findings
- The dataset contains 41 bills and laws, each verified on the legislature’s own page, as of 7 September 2026: 10 in force, 19 in committee, 2 enrolled, 2 introduced, 7 died and 1 veto sustained.
- The lane split is not one bill per lane. The 41 records carry 65 lane assignments: 18 for AI-generated evidence, 17 for courts and the judiciary, 14 for AI police reports and criminal-justice disclosure, 8 for attorneys’ use of AI, 5 for AI legal services and unauthorized practice, and 3 for legal aid and access to justice.
- California SB 574 passed both houses on 31 August 2026, was enrolled on 4 September 2026, and reaches attorneys and arbitrators through Business and Professions Code §6068.1, Code of Civil Procedure §128.7, Code of Civil Procedure §180 and Code of Civil Procedure §1282.1. Presentation to the Governor had not yet been recorded as of 7 September; once presented, the Governor has until 30 September 2026 under Cal. Const. art. IV §10(b)(2).
- Ten laws are in force across six jurisdictions as of 7 September 2026: California, Delaware, Louisiana, Maryland, Utah and Virginia.
- California SB 524 is in force from 1 January 2026 and requires AI-assisted official police reports to name the AI program, state that AI wrote all or part of the report and carry the preparer’s verification.
- Louisiana HB 178 / Act 250 is in force from 1 August 2025 and adds civil-procedure duties on false evidence, including evidence generated by AI or altered by any means.
- New York supplies 15 of the 41 records and no enacted law: its active cluster covers chatbot impersonation of licensed professions, civil filing affidavits, AI evidence reliability and police AI disclosure.
- Fourteen general AI bills were logged as excluded by the lane rule; examples include Texas TRAIGA, Colorado SB 24-205 and Connecticut SB 5. The NO FAKES Act is an example of the kind of digital-replica bill excluded unless a court-facing or legal-practice section is verified.
Why this question
United States AI bills are often grouped under consumer protection, privacy, deepfakes or agency governance. That grouping misses the smaller set that changes what happens in court or in legal practice. A bill about an AI assistant in a store is outside this report. A bill about an AI-drafted police report that reaches a prosecutor, a criminal defendant or an evidence hearing is inside it.
The difference matters because the operative section changes the reader. For a court administrator, the question is whether AI can create or alter a transcript. For a litigator, it is whether AI-created evidence is admissible or must be disclosed before trial. For a lawyer, it is whether a filing must disclose the use of a generative model and whether a citation has been personally verified. Each duty is traceable to a bill page and an official legislature page below.
The dataset also sits next to the live regulation record. Court rules and administrative orders are already tracked at /regulation/us-state, individual federal standing orders are discussed in /analysis/us-federal-court-ai-standing-orders, and courtroom misuse is tracked at /tracker. This report records the statutory layer only.
Method and data
This report uses version 0.1.3 of the SafeLegalAI legislation dataset, built on 7 September 2026. The public pages are /regulation/legislation, one page per bill at /regulation/legislation/<bill>, the JSON export at /regulation/legislation.json, and the GitHub release. The dataset contains 41 bills and laws, each verified on the legislature’s own page or GovInfo for federal bills. The site’s regulation taxonomy and editorial standard sit at /regulation#categories and /editorial-standards#regulation.
LegiScan, whose API documentation identifies the API as CC BY 4.0, and Open States were discovery sources. No bill was included from either lead alone. Every included bill was checked against the legislature’s own page or GovInfo for federal bills, and the passages quoted here come from public-domain operative text. Status is a dated snapshot, not a prediction. Bills move.
The lane rule is section-level. A bill is in only if an operative section regulates AI-generated or manipulated evidence in proceedings, AI use by courts or judicial agencies, AI legal services or unauthorized practice, attorneys’ AI use, AI-drafted police reports or criminal-justice disclosure, legal aid and access to justice, or court records for AI training. Broad AI governance, employment, lending, election deepfake, content-labeling and digital-replica bills are outside unless a court-facing or legal-practice section is verified.
Known limits are specific. The dataset covers 13 jurisdictions with verified records, not all 50 states plus the District of Columbia. States absent from the table should not be read as a no-bill finding. Fourteen near-misses were logged outside the count; six were retrieval failures and eight were outside the lane or lacked an AI-specific in-lane provision. Utah SB 180 is one of those near-misses because the official Utah text and status retrieval repeatedly timed out, so it is not counted or quoted here despite being a known law-enforcement AI lead.
The current enrolled California bill: SB 574
California SB 574 answers the legal-practice question directly. As enrolled, it would require attorneys using generative AI to keep nonpublic inputs out of unrestricted systems, verify every case and statutory citation, correct erroneous or hallucinated output and disclose use to the court for all documents submitted. The enrolled bill also changes Rule 11-style certification and sets limits for arbitrators.
| Item | Record |
|---|---|
| Register page | SB 574 |
| Official page | California bill history |
| Status | Passed both houses on 31 Aug 2026; enrolled on 4 Sep 2026; presentation to the Governor not yet recorded as of 7 Sep 2026 |
| Lanes | Attorneys’ use of AI; Courts and the judiciary |
| Action to watch | Once presented, Governor’s decision due by 30 Sep 2026 under Cal. Const. art. IV §10(b)(2) |
The operative §6068.1(a) in the dataset reads:
“An attorney shall not delegate the practice of law to generative artificial intelligence.”
The same operative text then requires confidentiality controls for inputs, reasonable steps to verify output accuracy, including “the accuracy of all case and statutory citations”, correction of erroneous or hallucinated output and disclosure to the court for documents submitted to the court. The bill would also change Code of Civil Procedure §128.7: no filed paper may contain citations the responsible attorney has not personally verified. For arbitrators, the enrolled text would prohibit delegating decision-making to generative AI and prohibit reliance on AI-generated information outside the record without prior disclosure.
California has already adopted a judiciary-side rule in the Regulation Tracker. Rule of Court 10.430 and Standard 10.80, listed at /regulation/us-state, require courts permitting generative AI use by judicial officers or staff to adopt written policies. SB 574 would sit on the practitioner and arbitrator side of that state record.
The laws already in force
The enacted set is small and concrete. Ten records are in force as of 7 September 2026. They do four things: regulate police and criminal-justice AI material, set evidence handling rules, create court or agency bodies, and extend consumer or professional-service AI duties into legal services. No enacted law is counted only because it mentions courts in findings.
| Jurisdiction | Bill | Effective | What it does | Primary |
|---|---|---|---|---|
| Delaware | HB 333 | 17 Jul 2024 | Creates the Delaware Artificial Intelligence Commission and directs it to recommend AI processes, principles and guidelines for executive, legislative and judicial agencies. | primary |
| Maryland | HB 338 | 1 Oct 2024 | House companion on law-enforcement facial-recognition results in criminal and delinquency proceedings. | primary |
| Maryland | SB 182 | 1 Oct 2024 | Regulates law-enforcement facial-recognition technology in criminal investigations and proceedings. | primary |
| Utah | SB 226 | 7 May 2025 | Applies high-risk generative-AI disclosure and occupation-compliance duties to regulated occupations, including licensed professional services. | primary |
| Maryland | SB 655 | 1 Jul 2025 | Creates an Artificial Intelligence Evidence Clinic Pilot Program in the Administrative Office of the Courts. | primary |
| Virginia | HB 1642 | 1 Jul 2025 | Requires human decision-maker involvement for listed criminal-justice decisions that use AI-based tools. | primary |
| Virginia | HB 2692 | 1 Jul 2025 | Bars AI-created inauthentic replica documents and other false statements in custodial interrogation of children, with an inadmissibility consequence. | primary |
| Louisiana | HB 178 | 1 Aug 2025 | Adds duties and pretrial procedures for false evidence, including evidence generated by AI or altered by any means. | primary |
| California | AB 316 | 1 Jan 2026 | Bars a defendant that developed, modified or used AI from asserting that AI autonomously caused the harm alleged. | primary |
| California | SB 524 | 1 Jan 2026 | Requires policies, disclaimers, program identification, verification and draft retention for AI-assisted official law-enforcement reports. | primary |
California SB 524 is the enacted AI-police-report law in force. Its operative text requires an official report generated partly or fully by AI to “identify every specific artificial intelligence program used” and state: “This report was written either fully or in part using artificial intelligence.” It also requires the preparer’s signature verifying review and factual accuracy.
Louisiana HB 178, enacted as Act 250, is the enacted evidence statute in the set. Its operative text says: “An attorney shall not knowingly encourage or produce false evidence, including evidence that is artificially generated or altered by any means”; the same text directs disclosure if a court determines that evidence has been falsified, including by AI generation or alteration.
Maryland’s enacted facial-recognition pair, SB 182 and HB 338, reaches evidence at the point of use in a case. The SB 182 operative text provides that facial-recognition results “may not be introduced in a criminal proceeding or in a delinquency proceeding”, subject to limited warrant and preliminary-hearing uses, and may not be the sole basis for probable cause or positive identification.
Maryland SB 655 reaches courts from the support side. It creates an Artificial Intelligence Evidence Clinic Pilot Program in the Administrative Office of the Courts. The operative text says the program provides expert testimony on the authenticity of electronic evidence that a court determines may have been created or altered using AI.
Virginia HB 1642 and HB 2692 are criminal-justice rules. HB 1642 says decisions on pretrial detention or release, prosecution, adjudication, sentencing, probation, parole, correctional supervision or rehabilitation “shall be made by the judicial officer or other person charged” and not without a human decision-maker. HB 2692 says a law-enforcement officer may not knowingly use false statements or AI-created inauthentic replica documents to secure a child’s cooperation, confession or conviction; a violating statement is inadmissible unless the Commonwealth proves waiver.
Utah SB 226 is narrower for this report than for general AI coverage. The counted provision is its regulated-occupation clause. The operative text says an individual providing services in a regulated occupation must disclose high-risk generative-AI interaction and “comply with all requirements of the regulated occupation when providing services through generative artificial intelligence.” Utah SB 180, by contrast, is not counted because the official text was not verified for this dataset.
The vetoed California evidence bill and other bills that stopped
Eight records have stopped advancing in the dataset: seven died and one has a sustained veto. The vetoed bill is California SB 11, which would have directed the Judicial Council to write rules for AI-generated or manipulated evidence. It is still useful because it shows the same policy route that California then pursued through SB 574 and court rules.
| Date | Jurisdiction | Bill | Status | Last action | Lanes | Primary |
|---|---|---|---|---|---|---|
| 2 Mar 2026 | California | SB 11 | Veto sustained | Veto sustained. | AI-generated evidence, Courts and the judiciary | primary |
| 7 Apr 2025 | Maryland | HB 966 | Died | Favorable Report by Judicial Proceedings; Withdrawn. | AI-generated evidence, Courts and the judiciary, Legal aid and access to justice | primary |
| 21 Feb 2025 | Kentucky | HB 498 | Died | to Judiciary (H) | Courts and the judiciary | primary |
| 5 Feb 2025 | Virginia | HB 2433 | Died | Left in Public Safety. | AI police reports and criminal-justice disclosure, AI-generated evidence | primary |
| 30 Nov 2024 | California | SB 970 | Died | From committee without further action. | AI-generated evidence, Courts and the judiciary | primary |
| 30 Nov 2024 | California | AB 2811 | Died | From committee without further action. | Attorneys’ use of AI, Courts and the judiciary | primary |
| 15 Feb 2024 | Congress | H.R. 7394 | Died | Introduced; referred to the House Committees on the Judiciary and Science, Space, and Technology. | AI-generated evidence, AI police reports and criminal-justice disclosure | primary |
| 22 Jan 2024 | Washington | SB 6073 | Died | Public hearing in the Senate Committee on Law & Justice at 10:30 AM. | Attorneys’ use of AI, Courts and the judiciary | primary |
California SB 11 would have added Evidence Code §1425. Its operative text required the Judicial Council, by 1 January 2027, to “review the impact of artificial intelligence on the admissibility of proffered evidence” and develop rules to help courts assess claims that evidence was generated or manipulated by AI. The Governor vetoed it in 2025 and the veto was sustained on 2 March 2026.
Washington SB 6073 was an early filing-disclosure bill. Its operative text required a party, attorney or pro se party using a generative AI language model for legal research or drafting Washington court filings to disclose the tool and certify that “each and every citation to the law or the record” had been verified as accurate. The bill received a public hearing on 22 January 2024 and then died.
Federal H.R. 7394, the Justice in Forensic Algorithms Act, died with the 118th Congress. Its operative text would have given criminal defendants access to reports, executable copies and source code for computational forensic software and made software-derived evidence admissible only after testing conditions were met.
New York’s cluster
New York accounts for 15 records because the same legislative idea appears across sessions, chambers and bill numbers. None is enacted in the dataset. The cluster divides into four lanes: chatbot liability where a bot impersonates an attorney or other licensed professional, civil filing affidavits for AI-drafted papers, criminal AI evidence reliability, and police AI inventories or disclosure.
| Group | Bills | Operative point | Status as of 7 Sep 2026 |
|---|---|---|---|
| Chatbot impersonation and licensed professions | A06545, S07263, S10675 | Operators may not knowingly permit chatbots to impersonate attorneys or other covered practitioners and may not disclaim liability through a non-human notice. | All in committee; latest action 5 Aug 2026 on S10675. |
| Civil filing affidavits | S09640, S02698, A08546, S09794 | AI-drafted civil papers would attach an affidavit disclosing use and certifying human review and verification. | All in committee; latest action 5 Jun 2026 on S09794. |
| Criminal AI evidence rules | A08110, S08390, A01338, A10625 | AI-created or AI-processed evidence would need independent support or reliability proof; A10625 would bar AI-generated outputs from court and legal decisions. | All in committee. |
| Police AI inventories and reports | A09253, S10425, A07172, S10574 | Covered AI in criminal investigations would be inventoried, disclosed in police reports or governed by DCJS protocol, including AI-generated report disclaimers. | All in committee; latest Assembly report 3 Jun 2026. |
The chatbot bills are unauthorized-practice bills because the covered-profession definition includes attorneys under Judiciary Law article 15. The A06545C operative text says an operator may not knowingly permit a chatbot to impersonate a covered professional in a manner that, if done by a natural person, would violate the article prohibiting the practice or appearance as an attorney without admission and registration. It also says an operator may not waive liability by telling consumers the chatbot is non-human.
The civil filing bills use a different mechanism. The S09794 operative text says any paper or file drafted with generative AI assistance must attach “a separate affidavit disclosing such use” and certify that a human has reviewed the source material and verified that the AI-generated content is accurate. The same language appears in earlier Senate and Assembly records.
New York’s police-report proposals go further than a public inventory. A09253A and S10425 define covered AI to include AI used to generate investigative leads and generative AI used to write police reports or other criminal-investigation records. The operative text says any covered-AI use in a criminal investigation “shall be disclosed in a police report” and that a police report created partly with generative AI must include an AI-content disclaimer.
New York also has a court-side record outside this legislation dataset. The Regulation Tracker lists the Unified Court System’s interim AI policy at /regulation/us-state, while this report lists only pending statutory proposals.
Federal bills and court records
The federal records are two live 119th Congress companions and one dead 118th Congress forensic-software bill. The live bills do not regulate lawyers’ filings or evidence admissibility directly. They would create a task force on AI speech-to-text and automatic speech-recognition technology in the United States judicial system and require a report to Congress.
| Congress | Bills | Lane | Status | Operative focus | Primary |
|---|---|---|---|---|---|
| 119th | H.R. 7997, S. 4154 | Courts and the judiciary | Introduced 19 Mar 2026 | Task force on AI speech-to-text and automatic speech-recognition technology in the U.S. judicial system. | H.R. 7997, S. 4154 |
| 118th | H.R. 7394 | AI-generated evidence; AI police reports and criminal-justice disclosure | Died 15 Feb 2024 | Defendant access and testing conditions for computational forensic software. | H.R. 7394 |
The H.R. 7997 and S. 4154 operative texts state the purpose as giving Congress an analysis and review of AI speech-to-text and automatic speech-recognition technology in the United States judicial system, plus recommendations. Their final report would address whether “a court record created or modified by AI speech-to-text technology or automatic speech recognition technology” affects the quality or accuracy of a legal-proceeding transcription.
Those bills sit beside, not inside, the federal standing-order record. Individual federal judges’ filing orders are collected in the separate SafeLegalAI analysis at /analysis/us-federal-court-ai-standing-orders. The two 119th Congress bills are about judicial-system records and court transcripts, not lawyer sanctions for fabricated citations.
Lane split and the excluded near-misses
The lane split shows why the dataset is not a general AI-bill tracker. Evidence and courts account for the largest coded groups, but the police-report lane is close because criminal investigations are the path by which AI-generated text, facial recognition and forensic software enter proceedings. Legal services and legal aid are smaller.
| Lane | Records |
|---|---|
| AI-generated evidence | 18 |
| Courts and the judiciary | 17 |
| AI police reports and criminal-justice disclosure | 14 |
| Attorneys’ use of AI | 8 |
| AI legal services and unauthorized practice | 5 |
| Legal aid and access to justice | 3 |
| Status | Records |
|---|---|
| In committee | 19 |
| In force | 10 |
| Died | 7 |
| Enrolled | 2 |
| Introduced | 2 |
| Veto sustained | 1 |
| Jurisdiction | Records |
|---|---|
| New York | 15 |
| California | 6 |
| Maryland | 4 |
| Congress | 3 |
| Virginia | 3 |
| Louisiana | 2 |
| West Virginia | 2 |
| Delaware | 1 |
| Kentucky | 1 |
| Oklahoma | 1 |
| Rhode Island | 1 |
| Utah | 1 |
| Washington | 1 |
Fourteen near-misses sit outside the count because the operative section did not reach the lane, or because a primary text could not be verified. That rule excludes Texas TRAIGA and the Colorado AI Act despite their general AI duties, because no strict courts, evidence, legal-services, attorney-use or police-report provision was verified. The NO FAKES Act is an example of the kind of digital-replica bill that stays outside unless a verified court-facing section is present.
| Bill | Jurisdiction | Reason |
|---|---|---|
| UT SB 180 (2025) | Utah | Known in-lane law-enforcement AI report bill, but primary Utah text and status retrieval repeatedly timed out, so no verified operative text was written. |
| UT HB 366 (2025) | Utah | Potential AI/evidence or public-safety lead; primary Utah text and status retrieval timed out and operative in-lane text was not verified. |
| WI AB 292 (2025) | Wisconsin | Official Wisconsin bill page verified metadata, but full bill text retrieval and Wayback text retrieval failed, so no operative text could be verified. |
| WI SB 295 (2025) | Wisconsin | Official Wisconsin bill page verified metadata, but full bill text retrieval and Wayback text retrieval failed, so no operative text could be verified. |
| CT SB 5 (2026) | Connecticut | Potential high-risk AI lead mentioning legal services, but primary Connecticut text and status retrieval timed out. |
| IL SB 3601 (2026) | Illinois | Professional AI Oversight Act lead was borderline for lawyers/legal services, and primary Illinois text/status retrieval timed out before scope could be verified. |
| NY S09832 (2025-2026) | New York | NCSL-style lead was inaccurate for this lane; official text concerns model-management contracts and digital replicas, not courts, legal services, evidence, attorneys, or police reports. |
| NY A03969 / S01274 (2025-2026) | New York | Official text addresses forensic science commission governance and forensic testing generally; no artificial-intelligence-specific operative court/evidence provision was found. |
| WV HB 3093 (2025) | West Virginia | Facial-recognition/privacy lead did not clearly regulate courts, evidence use, police reports, or legal services after primary review. |
| CA SB 69 (2025-2026) | California | AI transparency/consumer lead; primary review did not show a strict court, evidence, attorney, legal-services, or police-report provision. |
| CA SB 253 (2025-2026) | California | AI governance and public-sector lead; no verified in-lane operative provision for courts, evidence, legal services, attorney AI use, or police reports. |
| LA HCR 66 (2025) | Louisiana | Artificial intelligence study/resolution lead, but not specific enough to courts, evidence, legal services, attorneys, police reports, or court-records training. |
| CO SB 24-205 (2024) | Colorado | High-risk AI consumer-protection law; primary lane review did not identify courts, legal services, attorney AI use, police reports, or AI evidence provisions. |
| TX SB 2795 / TRAIGA (2025) | Texas | Broad responsible-AI governance lead; no strict in-lane operative provision was verified for this dataset. |
What to watch
| Date | Item |
|---|---|
| 30 Sep 2026 | California SB 574 passed both houses on 31 Aug 2026 and was enrolled on 4 Sep 2026; presentation to the Governor had not yet been recorded as of 7 Sep, and once presented the Governor has until 30 Sep under Cal. Const. art. IV §10(b)(2). |
| 4 Jun 2026 | New York’s official 2026 legislative session calendar lists this as the final session day; the New York cluster remains in committee as of 7 Sep 2026. |
| 19 Mar 2026 | H.R. 7997 was introduced and referred to the House Judiciary Committee; committee referral remains the recorded status as of 7 Sep 2026. |
Appendix A — data tables
The full 41-record table below uses the same lane and status names as the public legislation pages.
| Jurisdiction | Bill | Status date | Status | Lanes | Effective | Primary |
|---|---|---|---|---|---|---|
| California | SB 574 | 4 Sep 2026 | Enrolled; presentation not yet recorded | Attorneys’ use of AI, Courts and the judiciary | primary | |
| New York | S10675 | 5 Aug 2026 | In committee | AI legal services and unauthorized practice | primary | |
| New York | S09794 | 5 Jun 2026 | In committee | Attorneys’ use of AI, Courts and the judiciary | primary | |
| New York | S07263 | 5 Jun 2026 | In committee | AI legal services and unauthorized practice | primary | |
| New York | A07172 | 3 Jun 2026 | In committee | AI police reports and criminal-justice disclosure, AI-generated evidence | primary | |
| New York | A06545 | 1 Jun 2026 | In committee | AI legal services and unauthorized practice | primary | |
| New York | S10574 | 26 May 2026 | In committee | AI police reports and criminal-justice disclosure, AI-generated evidence | primary | |
| New York | S10425 | 15 May 2026 | In committee | AI police reports and criminal-justice disclosure | primary | |
| Congress | S. 4154 | 19 Mar 2026 | Introduced | Courts and the judiciary | primary | |
| Congress | H.R. 7997 | 19 Mar 2026 | Introduced | Courts and the judiciary | primary | |
| New York | A09253 | 17 Mar 2026 | In committee | AI police reports and criminal-justice disclosure | primary | |
| California | SB 11 | 2 Mar 2026 | Veto sustained | AI-generated evidence, Courts and the judiciary | primary | |
| New York | A08546 | 7 Jan 2026 | In committee | Attorneys’ use of AI, Courts and the judiciary | primary | |
| New York | A01338 | 7 Jan 2026 | In committee | AI-generated evidence | primary | |
| Maryland | SB 655 | 16 Dec 2025 | In force | AI-generated evidence, Courts and the judiciary, Legal aid and access to justice | 1 Jul 2025 | primary |
| California | AB 316 | 13 Oct 2025 | In force | AI-generated evidence | 1 Jan 2026 | primary |
| California | SB 524 | 10 Oct 2025 | In force | AI police reports and criminal-justice disclosure | 1 Jan 2026 | primary |
| Louisiana | HB 178 | 17 Jun 2025 | In force | AI-generated evidence, Attorneys’ use of AI | 1 Aug 2025 | primary |
| New York | S02698 | 13 Jun 2025 | In committee | Attorneys’ use of AI, Courts and the judiciary | primary | |
| Louisiana | HR 272 | 11 Jun 2025 | Enrolled | Courts and the judiciary | PDF; archive | |
| Maryland | HB 966 | 7 Apr 2025 | Died | AI-generated evidence, Courts and the judiciary, Legal aid and access to justice | primary | |
| Virginia | HB 2692 | 2 Apr 2025 | In force | AI-generated evidence, AI police reports and criminal-justice disclosure | 1 Jul 2025 | primary |
| Virginia | HB 1642 | 2 Apr 2025 | In force | Courts and the judiciary, AI police reports and criminal-justice disclosure | 1 Jul 2025 | primary |
| Utah | SB 226 | 27 Mar 2025 | In force | AI legal services and unauthorized practice | 7 May 2025 | primary |
| Rhode Island | S 0627 | 7 Mar 2025 | In committee | AI legal services and unauthorized practice | primary | |
| West Virginia | SB 688 | 4 Mar 2025 | In committee | AI police reports and criminal-justice disclosure | primary | |
| Kentucky | HB 498 | 21 Feb 2025 | Died | Courts and the judiciary | primary | |
| West Virginia | HB 2431 | 17 Feb 2025 | In committee | AI police reports and criminal-justice disclosure | primary | |
| Oklahoma | HB 2016 | 12 Feb 2025 | In committee | AI-generated evidence | primary | |
| Virginia | HB 2433 | 5 Feb 2025 | Died | AI police reports and criminal-justice disclosure, AI-generated evidence | primary | |
| California | SB 970 | 30 Nov 2024 | Died | AI-generated evidence, Courts and the judiciary | primary | |
| California | AB 2811 | 30 Nov 2024 | Died | Attorneys’ use of AI, Courts and the judiciary | primary | |
| Delaware | HB 333 | 17 Jul 2024 | In force | Courts and the judiciary | 17 Jul 2024 | primary |
| New York | A10625 | 20 Jun 2024 | In committee | AI-generated evidence, AI police reports and criminal-justice disclosure, Legal aid and access to justice | primary | |
| New York | S09640 | 16 May 2024 | In committee | Attorneys’ use of AI, Courts and the judiciary | primary | |
| Maryland | SB 182 | 16 May 2024 | In force | AI-generated evidence, AI police reports and criminal-justice disclosure | 1 Oct 2024 | primary |
| Maryland | HB 338 | 16 May 2024 | In force | AI-generated evidence, AI police reports and criminal-justice disclosure | 1 Oct 2024 | primary |
| Congress | H.R. 7394 | 15 Feb 2024 | Died | AI-generated evidence, AI police reports and criminal-justice disclosure | primary | |
| New York | S08390 | 26 Jan 2024 | In committee | AI-generated evidence | primary | |
| Washington | SB 6073 | 22 Jan 2024 | Died | Attorneys’ use of AI, Courts and the judiciary | primary | |
| New York | A08110 | 3 Jan 2024 | In committee | AI-generated evidence | primary |
Sources
Primary documents are the 41 official legislature and GovInfo pages linked in Appendix A, each also linked from its bill page. The bills quoted in the body are California SB 574, California SB 524, Louisiana HB 178 / Act 250, Maryland SB 182, Maryland SB 655, Virginia HB 1642, Virginia HB 2692, Utah SB 226, New York A06545C, New York S09794, New York A09253A, California SB 11, Washington SB 6073, federal H.R. 7997, federal S. 4154 and federal H.R. 7394.
Dataset and method sources: SafeLegalAI legislation register /regulation/legislation, JSON export /regulation/legislation.json, GitHub release, California Constitution article IV LegInfo page, Louisiana HR 272 official PDF and archive, LegiScan API licence page archive, Open States data page open.pluralpolicy.com/data, Regulation Tracker /regulation/us-state, federal standing-orders analysis /analysis/us-federal-court-ai-standing-orders, and the incident /tracker.
Reuse this research
Quote it, cite it, forward it — CC BY 4.0 for the data and figures; the linked official documents are the record. Suggested citation and a link to this exact report: