[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"$f2ggwRtnUU4mZmE2K6AvYLbcN5_SE5Dt1EOYmoz72MXU":3,"$fm6Ky0r9HHO8PTXKBgXJeU4fCLm4z6RKRMXFRXG3f560":28},{"success":4,"data":5},true,{"id":6,"slug":7,"title":8,"excerpt":9,"content":10,"category":11,"tags":12,"author":18,"cover_image_url":19,"reading_time_minutes":20,"is_published":4,"published_at":21,"created_at":22,"updated_at":23,"author_avatar":19,"is_featured":24,"meta_title":25,"meta_description":26,"meta_keywords":27},"7fe36f5c-d576-4c73-8b94-02a294e0400c","colorado-ai-act-rewrite-what-survived","Colorado's AI Hiring Law Was Delayed, Sued Over, and Rewritten — Here's What Actually Survived","Colorado's AI Act was delayed twice, sued into a stay, and repealed before a single enforcement action. What SB 26-189 actually asks of employers using AI in hiring, what's still unsettled, and why the rewrite isn't a retreat.","\u003Cp>For two years, Colorado's SB 24-205 — the Colorado AI Act (CAIA) — was the reference point for anyone tracking US state law on AI in hiring. It was the first comprehensive state AI statute in the country, and its \"high-risk AI system\" definition named employment decisions specifically. Every compliance checklist for AI screening tools cited it. Vendors built impact-assessment templates around it.\u003C\u002Fp>\n\n\u003Cp>It never actually took effect. On 14 May 2026, Governor Jared Polis signed Senate Bill 26-189, which repeals CAIA outright and replaces it with a narrower law built around notice and disclosure rather than risk management. The rewrite followed two delays, a special legislative session that failed to fix the original bill, and a federal lawsuit that froze enforcement before the ink on any of it was dry. This piece walks through what happened, what the new law actually asks of an employer using AI in hiring, and what is still unresolved — keeping \u003Cstrong>what the law now says\u003C\u002Fstrong>, \u003Cstrong>what remains unsettled\u003C\u002Fstrong>, and \u003Cstrong>our own reading\u003C\u002Fstrong> clearly separated.\u003C\u002Fp>\n\n\u003Ch2>How we got here: two delays and a lawsuit\u003C\u002Fh2>\n\n\u003Cp>CAIA passed in 2024 with an original effective date of 1 February 2026. It never reached that date intact.\u003C\u002Fp>\n\n\u003Cp>Colorado's legislature called a special session in August 2025 to fix provisions employers and civil-liberties groups both objected to — employers found the duty-of-care and annual impact-assessment obligations unworkable at scale; advocacy groups argued the bill's protections were too weak. The session failed to produce a compromise. Rather than let the original law take effect as written, Governor Polis signed SB 25B-004 on 28 August 2025, pushing the effective date to 30 June 2026 so the legislature could try again in its next regular session.\u003C\u002Fp>\n\n\u003Cp>Before that second attempt could even reach a vote, the law was pulled into federal court. In April 2026, Elon Musk's AI company x.AI sued Colorado Attorney General Phil Weiser (\u003Cem>x.AI LLC v. Weiser\u003C\u002Fem>, D. Colo.) to block CAIA, arguing the statute was unconstitutionally vague, compelled speech in violation of the First Amendment, burdened interstate commerce under the Dormant Commerce Clause, and denied equal protection. On 27 April 2026, the U.S. District Court for the District of Colorado granted a joint motion staying enforcement: the order barred the Attorney General from initiating any enforcement action under CAIA — or under any successor law the legislature might enact that same session — until well after the AG completes formal rulemaking and a subsequent injunction motion is resolved.\u003C\u002Fp>\n\n\u003Cp>That stay was still in place when the legislature returned to the bill. The Colorado General Assembly passed SB 26-189 on 9 May 2026; Governor Polis signed it five days later. CAIA, in the form everyone had spent two years preparing for, was repealed before a single enforcement action was ever brought under it.\u003C\u002Fp>\n\n\u003Cp>The lawsuit itself didn't happen in a vacuum — there is a federal current behind it worth naming, because it explains why a state law with no enforcement history drew the US Department of Justice. On 11 December 2025, the White House issued Executive Order 14365, \"Ensuring a National Policy Framework for Artificial Intelligence,\" directing DOJ to stand up an AI Litigation Task Force aimed at challenging state AI laws on preemption, Commerce Clause and related grounds; the Attorney General formally established the Task Force by memorandum on 9 January 2026. When DOJ moved to intervene against CAIA in April, the Department's own account described it as the first time it had intervened in litigation challenging a state AI statute — and, as of this writing, joining someone else's lawsuit remains the Task Force's only substantive action. The federal posture changed considerably; what it has not yet done is nullify a single state AI-hiring requirement actually on the books. Illinois's notice law and California's automated-decision-system rules are both still in force.\u003C\u002Fp>\n\n\u003Ch2>What the new law actually requires\u003C\u002Fh2>\n\n\u003Cp>SB 26-189 drops CAIA's core compliance machinery. Gone are the mandatory duty of care, the algorithmic impact assessments, and the formal risk-management-program requirement that made CAIA the most demanding state AI statute on paper. In their place, the new law — framed around \"automated decision-making technology\" (ADMT) used in \"consequential decisions,\" which explicitly includes employment — sets out a narrower, three-part obligation for employers:\u003C\u002Fp>\n\n\u003Col>\n  \u003Cli>\u003Cstrong>A pre-use notice\u003C\u002Fstrong> telling applicants or employees, before an ADMT-influenced decision is made, that the technology is in use and what it is evaluating.\u003C\u002Fli>\n  \u003Cli>\u003Cstrong>A post-adverse-outcome disclosure and human-review right\u003C\u002Fstrong> — when an automated tool materially influences an adverse decision (a rejection, a non-promotion, a termination), the affected person must be told and given a path to have a human review the outcome.\u003C\u002Fli>\n  \u003Cli>\u003Cstrong>Recordkeeping\u003C\u002Fstrong> — employers must retain records related to covered ADMT use for at least three years.\u003C\u002Fli>\n\u003C\u002Fol>\n\n\u003Cp>None of this takes effect yet. SB 26-189's framework becomes operative on \u003Cstrong>1 January 2027\u003C\u002Fstrong>, and the Colorado Attorney General is directed to adopt implementing rules — defining exactly what \"materially influence\" means, with illustrative examples and presumptions — by that same date. Under the terms of the April stay in the x.AI litigation, Attorney General Weiser has stated the office will not enforce CAIA or its successor until that rulemaking is complete and the litigation's remaining procedural steps play out. Practically, that means Colorado now has an AI-in-hiring law on the books with real notice, disclosure and recordkeeping obligations, an announced effective date, and an Attorney General who has publicly committed not to enforce any of it until sometime after rulemaking closes — which itself isn't due until the effective date arrives.\u003C\u002Fp>\n\n\u003Ch2>What's still unsettled\u003C\u002Fh2>\n\n\u003Cp>Two things are genuinely open, and we're not going to pretend otherwise. First, the definition of \"materially influence\" — the trigger for the post-adverse-outcome disclosure obligation — doesn't exist yet in operative form; it's the rulemaking's job to write it, and how narrowly or broadly the AG draws that line will determine how much of an ordinary AI-assisted hiring workflow the disclosure duty actually reaches. Second, the x.AI litigation hasn't been resolved, only paused by consent — a stay tied to rulemaking completion is not a ruling on the constitutional claims, and there's no guarantee the case ends quietly once rules are adopted. Anyone advising \"SB 26-189 is settled law, plan around it exactly as written\" is getting ahead of the record.\u003C\u002Fp>\n\n\u003Ch2>Where this fits in the wider patchwork\u003C\u002Fh2>\n\n\u003Cp>Colorado isn't the only state moving on this. Illinois' HB 3773 amendment to the Illinois Human Rights Act has been live since 1 January 2026, requiring employers to notify applicants and employees when AI is used in hiring and other employment decisions — a narrower, disclosure-only model that, notably, is already the shape Colorado has now converged toward. California has been updating its Fair Employment and Housing Act regulations to clarify how existing civil-rights protections apply to automated hiring tools, and Texas's Responsible AI Governance Act adds transparency and governance expectations for AI used in \"high-impact\" settings including employment. None of these statutes are identical, and none of them is a substitute for reading Colorado's actual rulemaking once it lands — but the direction across states is consistent: less emphasis on employers building internal risk-assessment bureaucracies, more emphasis on telling the person on the other end of the decision that a machine was involved and giving them a way to ask a human to look again.\u003C\u002Fp>\n\n\u003Ch2>Our reading\u003C\u002Fh2>\n\n\u003Cp>The easy takeaway from \"the toughest state AI law in the country got repealed\" is that AI-in-hiring regulation is losing momentum. We don't read it that way. What got repealed was the part of CAIA that was hardest for anyone — regulators included — to operationalize: formal duty-of-care standards and annual impact assessments with no settled methodology behind them. What survived, and what Illinois already has live, is the part that is actually easy to build and hard to excuse not having: tell people when AI is involved in a decision about them, tell them when it goes against them, let them ask for a human look, and keep the record. That's a lower bar to clear than CAIA's original text, not a lower bar to take seriously — 1 January 2027 is closer than it looks once a legal team has to draft notice language, wire up a human-review workflow, and set retention policies that don't already exist.\u003C\u002Fp>\n\n\u003Cp>We built our AI screening interviews around exactly that shape of obligation before any of these statutes required it of us. Every NiceHire AI screening greeting identifies the interviewer as an AI — in every language we ship, on first attempt and on reconnect after a dropped session — because the greeting pipeline itself won't let a customized or templated greeting remove that disclosure; it can change the wording, not the fact of it. And every completed AI screening interview keeps a transcript against the application record, so there's something to look back at if a candidate — or a regulator — asks what was actually said. Neither of those things exists because a Colorado statute told us to build them; they exist because \"the candidate should know they're talking to a machine, and there should be a record of what happened\" struck us as the right default regardless of which state's legislature got there first. If your hiring stack doesn't already do both, SB 26-189's 1 January 2027 date is a reasonable forcing function — but it shouldn't have to be the only reason.\u003C\u002Fp>\n\n\u003Chr>\n\n\u003Cp>\u003Cem>NiceHire Team\u003C\u002Fem>\u003C\u002Fp>","Compliance",[13,14,15,16,17],"Colorado AI Act","SB 189","AI regulation","compliance","hiring law","NiceHire Team",null,8,"2026-08-14T01:35:29.971+00:00","2026-08-14T01:35:29.951986+00:00","2026-08-14T01:35:30.040636+00:00",false,"Colorado AI Act Rewrite: What SB 26-189 Keeps for Employers","Colorado's AI Act never took effect — delayed, stayed, then rewritten. What SB 26-189 requires of employers using AI in hiring, and what is still unsettled.","Colorado AI Act, SB 189, AI regulation, compliance, hiring law",{"success":4,"data":29},{"posts":30,"count":67,"hasMore":4},[31,42,54,56],{"id":32,"slug":33,"title":34,"excerpt":35,"category":11,"tags":36,"author":18,"cover_image_url":19,"reading_time_minutes":40,"published_at":41},"fb2f1a0d-d2c3-4900-841d-fdbc1c6aa155","us-state-ai-hiring-law-patchwork-2026","The 2026 US State AI-Hiring Patchwork: What's Actually in Force, and What Just Got Walked Back","Illinois, California, New York City, Texas and Colorado are running five different tests for AI in hiring — one frozen by a federal court, one just publicly called out for under-enforcement. A map of what's actually binding in August 2026.",[37,38,13,39,16],"AI hiring law","state regulation","Local Law 144",10,"2026-08-14T01:35:40.383+00:00",{"id":43,"slug":44,"title":45,"excerpt":46,"category":11,"tags":47,"author":18,"cover_image_url":19,"reading_time_minutes":52,"published_at":53},"aa7b4a36-a110-4b8f-b0b0-d5460e015a61","pay-transparency-2026-written-notice-wave","Pay Transparency's Second Wave: Why 2026 Is About Individual Notices, Not Just Job Ads","Pay transparency's second wave isn't about job ads. Rhode Island, Oregon and California now require written notices to the individual hire, and New York is closing the placeholder-range loophole. What changed, and what it costs to miss.",[48,16,49,50,51],"pay transparency","salary ranges","written notice","employment law",11,"2026-08-14T01:35:36.487+00:00",{"id":6,"slug":7,"title":8,"excerpt":9,"category":11,"tags":55,"author":18,"cover_image_url":19,"reading_time_minutes":20,"published_at":21},[13,14,15,16,17],{"id":57,"slug":58,"title":59,"excerpt":60,"category":11,"tags":61,"author":18,"cover_image_url":19,"reading_time_minutes":20,"published_at":66},"c5db61a1-5f56-4273-988b-7fcb628288e2","ai-hiring-lawsuits-old-laws-new-tools","Old Laws, New Tools: What Three AI-Hiring Lawsuits in 2026 Are Actually Testing","The AI-hiring lawsuits moving through US courts in 2026 aren't brought under AI-specific laws. They're brought under statutes from 1967 and 1970 and a state lie-detector law — and they share a fact pattern worth checking your own stack against.",[62,63,64,65,16],"AI hiring lawsuits","Mobley v. Workday","FCRA","ADEA","2026-08-14T01:35:26.242+00:00",7]