The Consequential AI Evidence Monitor
Evidence Duties Timeline
Dated duties that make evidence a condition of using consequential AI: oversight, documentation, logging, provenance, independent audit, insurer testing. Each row states its current status and names its source.
Monitor · Method · Transactions · Duties · What systems prove
How to read this timeline
A row appears only after the instrument has been read at the issuing body’s site, under the Monitor’s method. The status column is the point of the page: a duty that is stayed, delayed, amended or under rulemaking says so on its row, with a date. Where an instrument already has a full page in the Regulayer law library, the row links to it; the library’s laws, rules, orders and country collections carry the wider record.
The timeline
| Jurisdiction | Instrument | Evidence duty | Key dates | Status at last review | Source |
|---|---|---|---|---|---|
| EU | AI Act, Articles 9, 13, 14, 15 (Regulation (EU) 2024/1689) | Risk management, transparency to deployers, human oversight by natural persons, accuracy and robustness for high-risk AI. | Entered into force 1 August 2024; obligations apply in stages. | In force, staged application. | Instrument page |
| EU | AI Act, Article 12 | Automatic event logging over the lifetime of a high-risk AI system. | Entered into force 1 August 2024; obligations apply in stages. | In force, staged application. | Instrument page |
| EU | GDPR, Article 22 | A right not to be subject to solely automated decisions with legal or similarly significant effect, and a right to obtain human intervention. | In force. | In force. | Instrument page |
| Ireland | High Court Practice Direction HC142 | A prescribed declaration on affidavits, witness statements and expert reports addressing generative AI use, with independent human verification. Paragraph 13 rejects using another AI system as sufficient verification. | Issued 29 July 2026; in operation 1 September 2026. | In operation from 1 September 2026. | Instrument page |
| US, Florida | Rule of General Practice and Judicial Administration 2.515(d)(2) | Every signature on a filing represents that the legal authorities cited exist and are accurately cited, with sanctions written into the rule. The duty falls on the named human signer whether or not AI was involved. | Effective 15 June 2026. | In force. | Instrument page |
| US, Colorado | SB 26-189, automated decision-making technology in consequential decisions | Notice, disclosure and consumer rights where automated decision-making technology is used in consequential decisions. | Signed 14 May 2026 (Session Laws Chapter 131). | Signed into law. The operative dates of its duties and the enforcement posture of Colorado’s framework are the subject of rulemaking and reported litigation; those specifics enter this row when the underlying documents have been read at source. | Bill page · Instrument page |
| US, New York City | Local Law 144 | Annual independent bias audit of automated employment decision tools, published results, and notice to candidates. | In force. | In force. | Instrument page |
| UK | Solicitors Regulation Authority, warning notice on the misuse of AI | Regulator notice: solicitors must check what is put before courts, supervise AI-assisted work, and remain accountable for it, with disciplinary action available. | Published 17 August 2026. | Published, current. | The notice · Instrument page |
| Canada, Ontario | Tribunals Ontario, Practice Direction on the Use of Artificial Intelligence | Adjudicator-facing: tribunal members do not use AI to write decisions or analyse evidence; parties must verify AI-assisted material and remain responsible for its accuracy. | April 2025. | In effect. | Instrument page |
| EU, pharma | EU GMP Annex 22, artificial intelligence (draft) | Documentation and control expectations for AI in GMP manufacturing. | Consultation draft. | Draft, not in force. | Instrument page |
| Insurance | AIUC-1 certification with insurer-backed coverage (ElevenLabs deployment) | Coverage for AI voice agents conditioned on certification testing. The program describes more than 5,000 adversarial simulations across security, safety, reliability, data privacy and accountability; the covered party reports 5,835 technical tests across 14 risk categories. | Announced 12 February 2026. | Announced by the covered party. | The announcement |
Coverage notes
- The timeline is not exhaustive. It is weighted to instruments already covered at full depth in the Regulayer law library, and it grows as further instruments are read at source.
- The principal free case-law hosts for Canada and Australia refuse automated access. Quiet from those jurisdictions on this page is a limitation of access, not a finding.
- Laws in this area converge toward oversight, documentation, logging, accountability and evidence. No instrument on this page mandates any particular vendor’s architecture, and this page does not claim otherwise.
Status is the point. A static list of statutes goes stale the day it ships. This timeline exists to carry the part that changes: stays, amendments, delayed application dates, rulemaking and enforcement developments, each stated with a date when the underlying document has been read.
Last reviewed 27 August 2026. Information, not legal advice. Deadlines and citations should be confirmed with counsel for the relevant jurisdiction.