Laws compared
A cell records what a text requires of the parties it defines. It does not show that any developer falls within them, or complied.
What does each binding government instrument that addresses developers of frontier or general-purpose AI models say on the same questions? Each column is one instrument; each row is one question, worded so that it fits every instrument. The questions are Selekoa’s; an instrument’s own terms appear only in quotation marks. Some instruments address providers of services rather than developers, and the first row quotes the party each one defines.
Question Q09Whistleblowers
Does the instrument protect people who report risks or violations?
Criterion: A provision prohibiting retaliation against employees or others who disclose risks or violations, or requiring a channel for such disclosures. A general complaint mechanism for users is noted but does not meet the criterion. Partly if the protection applies only by reference to another law.
What the statuses mean
- Stated
- A passage of the instrument meets the question’s criterion as worded, however conditional or narrow its scope. The condition or scope is in the note.
- Stated, partly
- A passage has what the criterion asks for, but in a weaker form than the criterion’s own words: for example, a power (“may”) where the criterion asks for a duty (“shall”), or a duty without the deadline the question asks about.
- Not found in the recorded search
- No passage meeting the criterion was found. The cell lists the sections read, the search terms and every passage set aside, with the reason. It does not show that the instrument lacks such a provision elsewhere, or that no other law imposes one.
- Does not apply
- The question does not arise for this instrument, for the reason given (for example, there is no framework document whose publication could be asked about).
Four rules decide “partly” in borderline cases:
- An order to an agency to impose a duty counts as partly.
- Protection only by reference to another law counts as partly.
- “Without undue delay” counts as a stated deadline.
- Required security measures for model weights count as risk-management measures under Q03.
Every stated or partly cell quotes the instrument’s own text with its article, section or page; Chinese passages carry Selekoa’s working translation. Three cells give a paraphrase with its locator instead, and say so. A not-found cell lists the sections read, the terms searched and every passage set aside, with the reason.
Paraphrased with its locator, not quoted: quoting it would pass the source’s quotation limit, or the source’s length is not known.
Limits of this comparison
- A cell records what the text requires of the parties it defines. It does not show that any named developer falls within them, or that anyone complied.
- A not-found cell is bounded by its search: the sections read and the terms searched. Another provision, or another law, may still impose the duty.
- The Code of Practice is a compliance route, not a law. Executive Order 14110 is revoked, and the BIS rule’s rescission was announced; each column shows its standing.
- The table compares and does not rank. Statuses are not scores and are not to be summed.
- Biological and chemical risk provisions are recorded at policy level only.
- Dates keep their kind: adoption, signature, approval, publication, entry into force and application are different events. A computed date is marked as computed.
How we searched
The instruments started from those on the frameworks-and-law page. A column was kept where the instrument is binding, or is the AI Act’s stated means of compliance (the Code), and places duties on developers or providers of frontier or general-purpose AI models. Fifty other instruments on that page were set aside, each with its reason, in the research pack: proposed or vetoed texts, guidance and reports, voluntary commitments, binding texts that place no duty on developers, and one standard whose text was not retrieved (GB 45438-2025).
United Kingdom: within the searches made, no UK law or other binding instrument placing duties on developers of such models was located. This is not a finding that none exists. The 2024 King’s Speech announced an intent to legislate; the 2026 briefing notes contain none of the searched terms; the memoranda with developers state that they are not legally binding; and title searches of legislation.gov.uk for 2023 to 2026 found one instrument, which places a duty on the Information Commissioner.
Other US federal texts: as read, none places a binding duty on developers. Executive Order 14409 of 2 June 2026 provides for a voluntary access framework and says that its section 3 authorises no licensing requirement.
Not yet read: the Republic of Korea’s AI Basic Act, Japan’s AI Promotion Act (2025), China’s provisions on deep synthesis (2022) and on algorithm recommendation (2021), and Colorado SB 24-205 (2024). They are not in the dataset; a later pack starts with Korea.
Illinois: the act’s official website refused the preparer’s requests, and the Internet Archive reset the connections. The cross-family checker’s session saved the enacted text from an Internet Archive capture of the official PDF (2026-09-08), and the cells were drafted from that saved text. The two model families read the act independently and agreed on the status of 13 of 15 cells. The approval date rests on the checker’s copy of the official bill status record, which the preparer could not open; it is given as a paraphrase, labelled as that copy.
Texts were read in official copies: the EU texts through the Publications Office’s repository, because EUR-Lex returned empty responses; California SB 53 from the Legislative Counsel’s database file, because its website refused requests; New York’s chapters in the Senate print and the Assembly’s bill records, not the session laws; the Chinese measures in the gov.cn policy library. Six web pages served changed bytes between retrievals; every quotation still matched.
Roles: one model family (Claude) drafted the cells, audited 50 of them in a fresh context and adjudicated. A model from another family (GPT, through Codex) checked 122 items, answering 85 of them blind before seeing the drafted answers. Where it disagreed, its status was applied. One cell it disputed (Executive Order 14110, Q06) was settled as not found after its second pass. For Illinois the roles were reversed: the checker read the act first, blind, and the drafter then wrote all fifteen cells without seeing its answers. The maintainer decided the remaining questions, including the four rules for “partly” and Illinois Q05 (partly, as for California’s identical wording). No person has checked the cells against the texts.
European Union
EU AI Act
Stated, partly
Paraphrase: Article 87 applies Directive (EU) 2019/1937 to the reporting of infringements of the Regulation and to the protection of the persons who report them.
Article 87
Note: Article 87 applies Directive (EU) 2019/1937 to the reporting of infringements of this Regulation and the protection of reporting persons; the protections are set by that Directive, which is not read here. Article 89(2) gives downstream providers a right to lodge a complaint, a complaint mechanism that does not meet the criterion. Partly because the protection is applied by reference to another law, not stated in this Regulation. Search for the rest of the criterion: Articles 85 to 94; Regulation (EU) 2026/1744 Article 1. Terms searched: “retaliat”, “whistleblow”, “Directive \(EU\) 2019/1937”, “reporting persons”, “complaint”. Passages read and not counted: Article 89(2): Complaint right for downstream providers; a complaint mechanism, not protection of reporting persons. Article 85: Complaints to market surveillance authorities about AI systems.
Inspect sources and checks: EU AI Act, Q09 WhistleblowersEU Code of Practice, Safety and Security
Stated
“Signatories will not take any legal or technical retaliation against the independent external evaluators”
Measure 3.5, paragraph on independent external evaluators, sentence before conditions (1)–(5) (PDF p. 14)
“not retaliating in any form”
Measure 8.3, point (7) (PDF p. 25)
Note: The Measure 3.5 commitment covers independent external evaluators, for their testing and/or publication of findings, as long as they meet five listed conditions. Measure 8.3 lists non-retaliation against persons providing information about systemic risks to competent authorities, and a published whistleblower protection policy, as examples of indicators of a healthy risk culture rather than as separate commitments. Recital (d) states that Measure 8.3 is without prejudice to Directive (EU) 2019/1937 and Article 87 AI Act.
Inspect sources and checks: EU Code of Practice, Safety and Security, Q09 Whistleblowers
United States: states
California SB 53
Stated
“prevents a covered employee from disclosing, or retaliates against a covered employee for disclosing,”
SEC. 4, Lab. Code § 1107.1(a)
“a reasonable internal process through which a covered employee may anonymously disclose information”
SEC. 4, Lab. Code § 1107.1(e)(1)
Note: Protection covers “covered employees”, defined in Lab. Code § 1107(b) as employees responsible for assessing, managing or addressing risk of critical safety incidents, who disclose to the Attorney General, a federal authority or persons with authority over the matter information on a specific and substantial danger from a catastrophic risk or a violation of the Act. The internal anonymous process (§ 1107.1(e)) is a duty of large frontier developers; § 1107.1(d) requires notice of rights, and § 1107.1(f)–(h) provide attorney’s fees, a burden-of-proof rule and injunctive relief.
Inspect sources and checks: California SB 53, Q09 WhistleblowersNew York RAISE Act (As amended, Chapter 96 of 2026)
Not found in the recorded search
- Read
- Bill text §§ 1–5, including GBL §§ 1420–1429 as re-added.
- Terms searched
- “whistle|retaliat|employee”.
- Passages read and set aside
- GBL § 1422(1) as re-added, p. 5: A mechanism for frontier developers or members of the public to report critical safety incidents to the office; a general reporting channel, with no protection against retaliation.
Note: No provision protecting persons who report risks or violations was found in the bill text of S8828.
Open the search record: New York RAISE Act (As amended, Chapter 96 of 2026), Q09 WhistleblowersNew York RAISE Act (As signed, Chapter 699 of 2025)
Not found in the recorded search
- Read
- Whole print S6953-B, pp. 1–5 (§ 1 short title; § 2 adding GBL §§ 1420–1425; § 3 effect).
- Terms searched
- “whistle|retaliat|employee”.
- Passages read and set aside
- Proposed GBL § 1420(1)(d), p. 2: “protect employee or customer privacy” is a ground for redaction, not a reporting protection.
Note: No provision protecting persons who report risks or violations was found in print S6953-B.
Open the search record: New York RAISE Act (As signed, Chapter 699 of 2025), Q09 WhistleblowersIllinois AI Safety Measures Act
Stated
“prevents a covered employee from disclosing, or retaliates … against a covered employee for disclosing, information to the Agency, Attorney General, a federal authority”
Section 20(a), PDF pp. 19–20
“a reasonable internal process through which a covered employee may anonymously disclose information”
Section 20(e)(1), PDF p. 21
“An employer may not take retaliatory action against an employee for disclosing or threatening to disclose in good faith any violation of the Artificial Intelligence Safety Measures Act.”
Section 90, Whistleblower Act Sec. 15(e), PDF p. 55
Note: A frontier developer may not adopt or enforce a rule, policy or contract that prevents, or retaliate against, disclosure by a “covered employee” (an employee responsible for assessing, managing or addressing the risk of critical safety incidents, Section 5) who has reasonable cause to believe the information shows a specific and substantial danger from a catastrophic risk or a violation of the act (Section 20(a)); notice of rights is required (Section 20(d)); large frontier developers provide an anonymous internal process with monthly updates (Section 20(e)). Section 90 amends the Whistleblower Act to protect any employee who discloses a violation of the act in good faith. “…” in the quotation stands for the running footer at a PDF page break; no words of the act are omitted.
Inspect sources and checks: Illinois AI Safety Measures Act, Q09 Whistleblowers
China
These measures address providers of services, not developers as such.
China, generative AI services (2023)
Not found in the recorded search
- Read
- Articles 1–24 in full.
- Terms searched
- “举报|投诉|报复”, “打击报复”, “员工|职工|工作人员”.
- Passages read and set aside
- Article 15 (第十五条): Providers must set up complaint and reporting (投诉、举报) mechanisms for the public with published handling procedures and response times: a general complaint mechanism for users, which the criterion excludes. Article 18 (第十八条): Users who find a service non-compliant have the right to complain or report to the competent authorities; a user complaint right, with no anti-retaliation protection.
Note: No provision prohibiting retaliation against people who disclose risks or violations, or requiring a channel for such disclosures by employees or others, was found. A public complaint and reporting mechanism is present (Article 15) and is noted.
Open the search record: China, generative AI services (2023), Q09 WhistleblowersChina, AI content labelling (2025)
Not found in the recorded search
- Read
- Articles 1–14 in full.
- Terms searched
- “举报|投诉|报复”, “员工|职工|工作人员”.
Note: No anti-retaliation provision or disclosure channel was found.
Open the search record: China, AI content labelling (2025), Q09 Whistleblowers
Revoked, or rescission announced
Kept as part of the record, each with its standing shown.
US Executive Order 14110
Not found in the recorded search
- Read
- Whole cached text of the Federal Register printing (88 FR 75191–75226), with close reading of Secs. 3, 4.1, 4.2, 12 and 13 and the signature block.
- Terms searched
- “whistle”, “retaliat”, “disclos”, “employee”.
Note: No provision on retaliation or on a disclosure channel for employees or others was found.
Open the search record: US Executive Order 14110, Q09 WhistleblowersUS BIS AI diffusion rule
Not found in the recorded search
- Read
- Federal Register PDF, 90 FR 4544–4584: DATES, SUPPLEMENTARY INFORMATION sections III and IV and Rulemaking Requirements; regulatory text §§ 740.27, 742.6(a)(13) and (b)(14), Supplement No. 10 to part 748 in full, ECCN 4E091 entry; signature and filing lines.
- Terms searched
- “whistle”, “retaliat”, “disclos”, “employee”.
- Passages read and set aside
- Supplement No. 10 to part 748, para. 18(c), 90 FR 4567: Programme for “detecting, assessing, disclosing, and managing insider threats”; security of the facility, not protection of those who report.
Note: No provision on retaliation or on a disclosure channel for employees or others was found.
Open the search record: US BIS AI diffusion rule, Q09 Whistleblowers