Research notes: Future’s AI lab needs to count the editing 18 November 2026 The feature distinguishes the 16 November iMore statement’s plans from measured results. The About page carries an earlier 3 August update date; this does not independently establish the wording present on that date or an actual launch date. The linked current Future and AI HTML page, labelled updated 12 December 2024, is the policy authority used here. It prohibits generating new articles for publication, separately restricts drafting without human core reporting, and permits repurposing completed human-authored text with verification. No policy breach, hidden directive or company-wide rule change is alleged. Paul Douglas’s quotation is from Introducing iMore Labs, 16 November 2026, not a Tomorrow’s Publisher interview. His role is taken from the original page’s current author biography. The 50/20-minute and 15/40-minute calculation is explicitly hypothetical. It illustrates 70 versus 55 total minutes, not actual iMore time or costs. Future’s Advisor figures are company-reported claims about recommendations and engagement, not article-generation results. WAN-IFRA’s CMS examples are supplier statements, not a verified Future deployment or measured saving. The Guardian comparison is a written rule. The FT case study is qualitative, based on interviews in May-July 2025 and published 3 October 2026. Coauthor Liz Lohn worked at the FT during the research. Neither supplies a comparative labour-and-error result for iMore. Noah’s bounded enquiry covered publishing and newsroom technology over 18 November 2025 to 18 November 2026. A further four-lane discovery check was completed after the full first draft. Off-subject material and broad retrieval volumes were not used as adoption rates, sentiment measures or event probabilities. Automatic forecast timing limitation: the same cumulative two-title rollout event received 34% by 18 May 2027 and 29.4% by 18 November 2027. These point estimates are not monotonic. Both are retained unchanged below, but neither is used in the feature and no coherent rollout timing path is inferred. No number, date or target was changed or recalculated. The article uses two separate forecasts: effective corporate drafting authority and publication of a qualifying workflow evaluation. The evaluation forecast concerns public evidence appearing, not whether private gains exist. The specific withdrawal-event forecast is not used because it would not establish ordinary productivity or a representative failure rate, and no present iMore fabrication is alleged. Research note: Noah completed 38 of 40 planned research checks; 2 unfinished checks remain outside the conclusion; the detailed method and source trail are retained in the Claim audit. Full original automatic forecast register: future_drafting_beyond_lab_20270518 Between 2026-11-19 and 2027-05-18, will at least two Future-owned editorial titles other than iMore publish original news or feature articles substantially drafted with generative AI as part of their ordinary editorial production, excluding recommendation tools, syndicated material and articles published solely as AI demonstrations? YES: 34%; NO complement: 66%; uncertainty interval: [12, 60]%. Deadline: 2027-05-18 YES conditions: By 2027-05-18, at least two qualifying titles have each published such an article after 2026-11-18. Substantial drafting means generating original article prose, not merely transcription, translation, proofreading or headline suggestions. Publisher workflow accounts, article disclosures and attributable reporting in Press Gazette or journalism.co.uk may verify occurrence. NO conditions: By 2027-05-18, fewer than two qualifying titles have done so during the specified period. Missing AI labels alone do not establish nonoccurrence; resolution should consider publisher accounts and independent reporting. Run reference: 9591da8fb44560131cc2037de976c4042cd14dda486cdf8ba4cb968a2dd7f191 Origin: Noah final automatic quick forecast. Absence of a label, report or correction notice alone does not establish nonoccurrence, except where the defined event itself is publication of an evaluation. future_drafting_beyond_lab_20271118 Between 2026-11-19 and 2027-11-18, will at least two Future-owned editorial titles other than iMore publish original news or feature articles substantially drafted with generative AI as part of their ordinary editorial production, excluding recommendation tools, syndicated material and articles published solely as AI demonstrations? YES: 29.4%; NO complement: 70.6%; uncertainty interval: [12, 58]%. Deadline: 2027-11-18 YES conditions: By 2027-11-18, at least two qualifying titles have each published such an article after 2026-11-18. Substantial drafting means generating original article prose, not merely transcription, translation, proofreading or headline suggestions. Publisher workflow accounts, article disclosures and attributable reporting in Press Gazette or journalism.co.uk may verify occurrence. NO conditions: By 2027-11-18, fewer than two qualifying titles have done so during the specified period. Missing AI labels alone do not establish nonoccurrence; resolution should consider publisher accounts and independent reporting. Run reference: 09c1b18a236ab2853afa30a9bbf9bc41a7678187e97587eba067fb3084651928 Origin: Noah final automatic quick forecast. Absence of a label, report or correction notice alone does not establish nonoccurrence, except where the defined event itself is publication of an evaluation. future_drafting_authority After 2026-11-18 and by 2027-11-18, will Future put into effect a corporate editorial AI policy that expressly permits generative-AI drafting of original articles for publication beyond iMore Labs, with human editorial approval required before publication? YES: 35%; NO complement: 65%; uncertainty interval: [15, 65]%. Deadline: 2027-11-18 YES conditions: By 2027-11-18, a Future corporate editorial AI policy has taken effect expressly authorising the specified drafting beyond iMore Labs and requiring human editorial approval. Future's operative policy, supported where necessary by attributable company statements or Press Gazette reporting, can verify its scope and effective date. NO conditions: By 2027-11-18, no such policy has taken effect. A proposal, a lab-only exception, permission limited to recommendation tools, or informal use without the specified corporate authorisation does not qualify. Run reference: bd1c03fb2706f4178a84ab2eb0015faca55980576727912463a8f98fffb5c079 Origin: Noah final automatic quick forecast. Absence of a label, report or correction notice alone does not establish nonoccurrence, except where the defined event itself is publication of an evaluation. end_to_end_workflow_evidence Between 2026-11-19 and 2027-11-18, will a UK- or Europe-headquartered news publisher, or a research organisation studying one, publish a comparative evaluation of an implemented AI-assisted article-drafting workflow that reports both total human labour through publication and factual-error outcomes against a non-generative-AI drafting baseline? YES: 38.4%; NO complement: 61.6%; uncertainty interval: [18, 62]%. Deadline: 2027-11-18 YES conditions: By 2027-11-18, a publicly accessible evaluation published during the specified period compares implemented AI-assisted and non-generative-AI drafting workflows for comparable editorial tasks, reports human time including research, drafting, verification and editing through publication, and reports factual-error outcomes for both. Publisher research pages, Reuters Institute, WAN-IFRA or an accessible scholarly paper may supply the record; favourable results are not required. NO conditions: By 2027-11-18, no qualifying evaluation has been published. Draft-generation speed, output counts, plans, testimonials or evaluations omitting either end-to-end human labour or comparative factual-error outcomes do not qualify. This does not establish that underlying productivity gains are absent. Run reference: 1d66560f8c19b123256eb78f881a17203b0d322a1a6ae92848527a1869b84691 Origin: Noah final automatic quick forecast. Absence of a label, report or correction notice alone does not establish nonoccurrence, except where the defined event itself is publication of an evaluation. imore_fabrication_withdrawal Between 2026-11-19 and 2027-11-18, will iMore withdraw from ordinary reader access an article substantially drafted with generative AI because it contains a fabricated quotation, source or cited document? YES: 5%; NO complement: 95%; uncertainty interval: [1, 15]%. Deadline: 2027-11-18 YES conditions: By 2027-11-18, iMore has withdrawn a qualifying article during the specified period because of at least one fabricated quotation, source or cited document. Substantial drafting means generating original prose rather than proofreading or transcription. iMore correction or withdrawal records, archived article versions and attributable Press Gazette or journalism.co.uk reporting may establish the drafting method, defect and reason for withdrawal. NO conditions: By 2027-11-18, no qualifying withdrawal has occurred during the specified period. Routine corrections, broken links, temporary technical outages, unrelated removals and fabricated material in articles not substantially AI-drafted do not qualify. An absence of a correction notice alone is insufficient evidence of nonoccurrence. Run reference: fdab30634bc7f109cd051baa9d32a64c0b1025a7589bf9443f76da2d22cff98f Origin: Noah final automatic quick forecast. Absence of a label, report or correction notice alone does not establish nonoccurrence, except where the defined event itself is publication of an evaluation. Sources: Paul Douglas: Introducing iMore Labs, 16 November 2026 https://www.imore.com/apple/introducing-imore-labs Future and AI: current linked corporate guidance, page dated 12 December 2024 https://futureplc.com/future-and-ai/ About iMore Labs, marked updated 3 August 2026 https://www.imore.com/about Future Unveils Advisor AI, 2 July 2025 https://futureplc.com/news/future-unveils-advisor-ai/ Guardian editorial code and guidance, March 2026, page 12 https://uploads.guim.co.uk/2026/03/03/Editorial_Code_of_Practice_Guidelines_March2026.pdf Liz Lohn and Felix M. Simon: Financial Times AI transparency case study, 3 October 2026 https://link.springer.com/article/10.1007/s00146-026-03369-z WAN-IFRA webinar report: CMS platforms and embedded AI, 30 April 2026 https://wan-ifra.org/2026/04/cms-ai-newsroom-workflows-integration/