TOPIC INTELLIGENCE

Human Oversight

Compare how verified workplace-technology frameworks address human involvement, review and accountability without assuming that every framework provides the same rights.

Coverage scope

  • Human corroboration of automated outputs
  • Review and appeal provisions
  • Vendor documentation and employer responsibilities
  • Enacted safeguards versus proposed duties

CURRENT COVERAGE

9 records · 8 articles

Relevant jurisdictions: California, Colorado, New York, Texas, Vermont, Washington.

REGULATORY RECORDS

Effective — ADMT duties begin January 1, 2027regulation

CCPA Risk Assessment and Automated Decisionmaking Technology Regulations

Final CCPA regulations establish risk-assessment requirements for specified high-risk processing beginning in 2026 and notice, access, opt-out, and appeal provisions for covered ADMT used in significant decisions beginning January 1, 2027.

Related coverage: California Enacts New Workplace AI Rules: What SB 947 and AB 1883 Require →
Chaptered — operative July 1, 2027statute

California SB 947 — Employment: Automated Decision Systems

SB 947 limits employer reliance on automated decision systems in employee discipline and termination and adds human-corroboration, postuse-notice, and employee-data-description requirements when reliance is primary.

Related coverage: California SB 947: What Does Human Review Actually Require Before AI-Driven Discipline or Termination? →
Signed — operative January 1, 2027; rules proposedstatute

Colorado Automated Decision-Making Technology Act

Colorado repealed and reenacted its automated-decision framework. The replacement law applies to consequential decisions made on or after January 1, 2027, including covered employment decisions; the October 6, 2026 interim draft revises the August 11 proposed implementing rules and is not final.

Related coverage: If a staffing firm uses AI to screen applicants, who is responsible under Colorado’s proposed rules? →
Vetoedbill

California SB 7 — Employment: Automated Decision Systems

SB 7 proposed notice, access, correction, appeal, and human-oversight requirements for employer use of automated decision systems. The Governor vetoed the bill on October 13, 2025; it did not become law.

Related coverage: California AI Hiring Compliance in 2026: How FEHA and CCPA Automated Decision Rules Fit Together →
In Senate Labor Committeebill

New York Boundaries on Technology Act

The BOT Act proposes impact assessments, notice, restrictions on electronic monitoring and automated employment decision tools, data-access rights, and meaningful human oversight. The official Senate record currently lists S 10147 as active in the Senate Labor Committee.

Related coverage: State AI Employment Laws: What’s Law, What’s Pending, and What’s Guidance →
Referred to House committeebill

Vermont H.340 — Automated Decision Systems

H.340 proposes a broad framework for developers and deployers of automated decision systems used for consequential decisions, including employment decisions. The official status page lists the bill in the House Committee on Commerce and Economic Development and shows no enacted act.

Related coverage: State AI Employment Laws: What’s Law, What’s Pending, and What’s Guidance →
Retained in House Appropriationsbill

Washington HB 1672 — Technology Used by Employers in the Workplace

HB 1672 proposes restrictions and procedural protections for employer electronic monitoring and automated decision systems. After a substitute received a do-pass recommendation, the bill was referred to House Appropriations and retained in that status for the 2026 session.

Related coverage: State AI Employment Laws: What’s Law, What’s Pending, and What’s Guidance →
Did not become law in 2026bill

Connecticut SB 435 — Automated Decision Systems Protections for Employees

SB 435 proposed a broader employment automated-decision framework, including explanations, data correction, human review, bias audits, recordkeeping, collective-bargaining provisions, and a civil action. The official history ends with an April 20, 2026 favorable report and Senate calendar placement; the regular session adjourned May 6 without enactment.

Related coverage: Connecticut’s AI Employment Law Is Now Partly in Effect: What Employers Need to Know →

LATEST COVERAGE

Explainer

NY · CO · CA · IL · MD · MA · NJ

AI Bias Audits: One-Time Compliance or Continuous Monitoring?

Current law does not impose one universal monitoring model. New York City uses a rolling one-year audit rule; other regimes rely on discrimination law, notices, records, documentation, or human review—and governance programs often go further.