Check who the program covers
These examples concern government workforces with different rules and conditions. They do not create a single training requirement for every US employer. Confirm your organization’s role and applicable policy before choosing a course.
Announced access, enrollment and completion are different measures. This guide concerns people learning to use and govern AI; it does not describe rules for data used to train AI models.
Liquid Learn provides software for coordinating and tracking governance work. The official programs below are independent of Liquid Learn; links do not imply partnership, endorsement, course accreditation or an LMS integration.
See training rollout in the product workflowTexas
- Covered people
- Covered state and local employees and officials. Computer-use/access thresholds and exceptions apply; school districts have specific rules. Check DIR’s role table.
- Authority
- Texas Government Code §§ 2054.5191 and 2054.5193; DIR implementation guidance.
- Rollout / policy date
- Current DIR materials cover FY 2026–27; listed course certifications run through August 31, 2027.
- Cadence / trigger
- Annual training; entities certify compliance to DIR by August 31.
DIR’s free awareness course does not track learners or issue completion certificates. Employers document completion using their chosen method; individual training records are not submitted to DIR.
Organizational practice: Keep the covered population, selected certified course and completion evidence together. Liquid Learn is not represented as a DIR-certified training provider.
Sources checked
California
- Covered people
- Executive Branch State Entities and their personnel. This is not a general private-employer training mandate.
- Authority
- State Administrative Manual 4986.13; CDT Technology Letter 25-01.
- Rollout / policy date
- SAM section introduced February 2025; the technology letter gives February 20, 2025 as the policy effective date.
- Cadence / trigger
- Training necessary for duties and applicable mandatory GenAI training. No universal annual interval is specified in this section.
Entities must ensure necessary role-relevant training. Foundational learning before tool deployment and specialist learning for managers, procurement, project and technical roles are additional practices the policy says to consider.
Organizational practice: Separate required learning from recommended role pathways. CDT offers technical training and points general workforce/business leaders to CalLearns.
Sources checked
New York
- Covered people
- State workers; the AI Pro tool is offered to the 50-plus agencies/entities supported by ITS.
- Authority
- New York ITS workforce rollout announcement and AI Pro program condition.
- Rollout / policy date
- April 6, 2026: expansion from a 1,200-person pilot to offering training to more than 100,000 state employees.
- Cadence / trigger
- Responsible-AI training is required for agencies electing to use AI Pro. The announcement does not establish an annual refresher interval.
The rollout combines two-part InnovateUS learning with AI Pro. Announced access is not a count of completed learners.
Organizational practice: Track rollout readiness and completion separately. This program condition does not establish a blanket mandate for New York employers.
Sources checked
Pennsylvania
- Covered people
- Commonwealth employees participating in the approved generative-AI rollout.
- Authority
- Office of Administration program guidance and April 15, 2026 expansion update.
- Rollout / policy date
- The April update reported more than 3,000 employees using AI across 35 agencies, with an additional 6,500 enrolled in required learning.
- Cadence / trigger
- Training is a condition of use in this rollout. No universal annual interval is stated in the reviewed announcement.
Approved tools, InnovateUS learning and governance oversight accompany expansion. Enrollment is not completion; the reported figures are program activity, not Liquid Learn customer results.
Organizational practice: Review learner status alongside approved uses and oversight responsibilities. Completion alone does not grant software access.
Sources checked
A role-based approach for private employers
NIST’s voluntary AI Risk Management Framework offers a useful organizing approach. GOVERN 2.2 addresses learning suited to people’s responsibilities, policies and risk-management duties, including escalation and significant changes. Read NIST GOVERN 2.2.
NIST states that AI RMF 1.0 is under revision. Using the framework does not replace a determination of applicable law or sector duties. Check the framework’s status.
Suggested organizational approach: cover AI limitations, approved uses, privacy and security, checking outputs, bias and human oversight, reporting problems, and practice suited to each role. This is an editorial suggestion, not a state-approved syllabus. Set refreshers around changed duties, tools, policies and guidance.
Keep EU obligations in their own jurisdictional context. Read the EU AI Act guide separately from these US workforce programs.
Keep the rollout accountable
- Define the people and owner. Identify affected teams, AI uses and oversight responsibilities.
- Select and document the learning. Record the provider, course version and why it fits the role. Confirm recognition with the relevant authority.
- Coordinate delivery. Set deadlines and communicate the rollout. Confirm how your existing learning systems will participate.
- Review evidence and gaps. Distinguish invitations, enrollment, progress and completed learning. Assign follow-up for missing records.
- Revisit the requirement. Review after changes to tools, duties, approved uses or policy. Retain the records required by your organization.
Liquid Learn supports requirements with owners, targeted course/assessment items, assignment and completion views, and evidence references. Your team remains responsible for course suitability, communications, follow-up and any required government reporting.
Request a product demoSources and editorial review
Editorial owner: Fluid Business — Liquid Learn editorial team. Last checked September 26, 2026; next editorial review due October 26, 2026, or sooner if an official source changes. Report a correction.
This is a focused source review, not continuous regulatory monitoring or a 50-state inventory. Other states are unreviewed here. Washington is omitted because its operative training-policy document was not reliably verified; that does not mean no requirements exist.
For broader governance questions, return to the US governance guide.
Bring your training rollout to a product demo
Explore training requirements, learner assignments, completion gaps and evidence in Liquid Learn. Connect those records to the owners and AI services they support.
Educational information, not legal advice. Applicability depends on jurisdiction, organizational role and use case. Confirm your obligations with qualified counsel. Liquid Learn supports governance work; using it does not certify or guarantee compliance.