Every AI-touched vertical becomes a regulated one; this newsletter tracks the arc in motion. Sector trends free; deeper methodology, cross-sector overlays, and quarterly rollups behind the $49/mo unlock. Manifesto here.
Every AI-touched vertical eventually becomes a regulated one. The manifesto that anchors this newsletter walks through why the arc is compressing to 18-36 months, why the plaintiffs' bar functions as the arc's referee during the growth-through-pain phase, and why the eventual consumer outcome across every industry that has walked this arc before is a raised floor rather than a cap. This monthly issue is the arc-tracker: a descriptive snapshot of where the arc appears to be moving fastest right now.
Regulation of a sector's AI use follows adoption by three to five years. The published-and-covered sectors get the regulatory ink first — healthcare, banking, insurance, education, financial services. Everyone else follows. Public SEC filings are the earliest signal for where sector-specific AI rulemaking is likely to arrive next, because filers describe AI-adoption in disclosure language 12-36 months before regulators codify it.
This issue tracks four sector clusters where AI-related disclosure density has risen materially inside a 20-day sampling window. The clusters are chosen for the descriptive strength of the pattern, not for any prediction about specific corporate behavior. Methodology detail, cross-sector overlay analysis, and the quarterly rollup sit behind the subscription tier.
"artificial intelligence", "generative AI", "large language model", "AI hallucination", "AI governance", "AI-generated", "algorithmic bias", "model drift", "AI-related litigation", "AI-related regulatory", "AI agent", and "machine learning" AND "regulatory" — across 8-K, 10-Q, 10-K and DEF 14A filings between 2026-07-30 and 2026-08-18. Fourteen trading days, 802 filings from 666 unique filers surfaced by keyword. Sector-cluster assignment is a manual review pass by the author, not an automated industry-code join: each filing in a named cluster was read and classified by hand, which is why cluster membership does not always track a filer's primary SIC code. Cluster counts are counts of filings.Investor-owned utilities, gas distributors, and next-generation nuclear and geothermal filers surfaced AI-mentioning disclosure language in 16 filings during the sampling window. Language varied from routine risk-factor references to specific machine-learning-in-regulatory-context statements. Federal energy regulation (FERC) and reliability standards (NERC) do not yet have sector-specific AI-governance rules. State PUC AI-rulemaking activity is at the pre-notice-of-proposed-rulemaking stage in most jurisdictions.
Real-estate investment trusts, homebuilders, self-storage operators, and commercial real-estate advisors disclosed AI-adoption language in 10 filings during the sampling window. Notable pattern: four sub-fund filers under one sponsor used identical machine-learning-in-regulatory-context language across the same week, indicating either a shared disclosure template or a shared underlying operational reality worth watching in future cycles.
Sidewalk-delivery robotics, restaurant and hospitality service robotics, and industrial automation semiconductor filers disclosed generative-AI or AI-adoption language during the sampling window. Product-liability doctrine and product-safety disclosure norms apply to physical-world robotics deployment independent of any AI-specific rulemaking; disclosure practice around AI components of physical systems is still evolving.
AI-based weapons-detection, security-monitoring, and identity-verification filers disclosed AI-adoption language during the sampling window. Federal FTC enforcement activity and California and New York legislative activity around AI in security and identity are further along here than in the other three clusters. The FTC has not promulgated AI-specific rules in this area; it has proceeded by enforcement action and consent order. Disclosure practice is evolving quickly in response.
Filers with AI-adoption disclosure language but not yet cluster-density: Performance Food Group (food distribution 10-K, generative-AI language), Trade Desk and DoubleVerify (adtech 10-Q, machine-learning-in-regulatory-context language), Groupon (8-K), e.l.f. Beauty (10-Q, generative-AI marketing language), AIxCrypto Holdings, Universal Token (crypto AI-agent language).
Shipping with Issue #1 (33 pages): the full roster of all 802 matched filings across 666 unique filers, every one with its accession number and a direct SEC.gov link; the exact screening specification; per-day disclosure density; form and search-string breakdowns; repeat-filer analysis; and the co-filing detection behind the sub-fund boilerplate observation. Every figure is computed from the screening data by a script that ships with it, so you can reproduce the document and check the work.
Not in Issue #1: the regulator-comment tracking, state rulemaking calendars, and historical rule-cycle comparisons listed in the section boxes above. Those require primary docket and legislative research; they go in as they are written rather than filled with generated text. The attachment states the same scope boundary in full.