CAWS
Training

Training that ends in a decision workers can act on.

Four courses, from a short awareness session to applied work on a system active in their workplace. Participants learn to find AI in a work task, describe the hazard specifically enough that someone can act on it, and distinguish a real control from just shifting responsibility to an individual.

Talk about a delivery
The courses

Four tiers, one method.

Courses run for single-role groups by default, since sessions cover reporting and recourse. Delivered in person. Pricing and scoping on request.

Tier 1 · Awareness 120 minutes

AI in the Workplace

A short session for anyone who works alongside a system that decides, sorts, scores, schedules, or writes. No technical background needed.

Who it is forStaff at any level in public agencies, private employers, nonprofits, and unions. Office and professional roles, frontline and field roles, supervisors, and managers.
Participants leave able to
  • Locate AI in a work task, including where it sits inside another system and is not visible
  • Distinguish a real control from one that shifts responsibility to an individual
  • Document an incident and know when to stop or escalate
90-minute version available Workbook they keep 30-day follow-up
Tier 2 · Applied One day

AI in the Workplace: Analyzing Hazards and Controls

Participants run a full AI job hazard analysis on a system from their own workplace and leave with a plan they can act on.

Who it is forProgram and operations managers, compliance and HR staff, team leads, and safety and health professionals. Also worker representatives, trainers, and stewards.
Participants leave able to
  • Map a work system and identify where AI enabled decisions enter it
  • Rank hazards by consequence, exposure, and detectability
  • Design a control that acts on the system, or state plainly what prevents one
Completed AI job hazard analysis Escalation route 30, 60, 90 day plan
Tier 3 · Specialist Two days

AI in the Workplace: Hazards, Controls, and Assurance

For practitioners responsible for how AI enabled systems are assessed, controlled, and governed across an organization.

Who it is forRisk and governance leads, program directors, and compliance officers, along with occupational and environmental health and safety professionals and industrial hygienists. A working background in risk, governance, or safety is expected.
Participants leave able to
  • Build and maintain an AI hazard register covering several systems at once
  • Write controls specific enough to be adopted into workplace governance, procurement conditions, or contract language
  • Separate controls the organization can implement through work design from controls that require the vendor to change the software
Hazard register Control architecture Verification plan

Modules covering machine safety, interlocks, and control validation are delivered with a subject matter specialist alongside the CAWS facilitator.

Tier 4 · Applied engagement Scoped to the system

Implementation Lab

CAWS runs the AI job hazard analysis on one of your own systems, already in service or anticipated for deployment, with the people who operate it and the people responsible for it in the room together.

Who it is forOrganizations with an internal function that owns risk, governance, compliance, or safety. Tier 2 or Tier 3 completion helps.
Participants leave with
  • A hazard register with exposure pathways, consequence, and detectability
  • A control specification that can be built, bought, or required of a vendor
  • A verification plan with the evidence that would show each control is working
On site Three to five engagement days 30 and 90 day follow-up

CAWS is not affiliated with any vendor and does not promote, sell, or certify any AI product or tool. CAWS does not act as an auditor and does not approve or clear any system.

How this fits the rest of the work

Training is also a listening site.

What participants surface about the systems running in their workplace feeds the evaluation criteria and the open research CAWS publishes. Engagements include a follow-up conversation, booked before the session is delivered.

The method

An AI job hazard analysis, extending the job hazard analysis already used in safety work. It locates AI in a work system, identifies who is exposed, and selects controls using the hierarchy of controls. Notice, explanation, challenge, and correction are treated as controls to be designed in, not as policy language.

Availability

Open for pilot delivery through fall 2026.

Pilot engagements

A delivery of AI in the Workplace, or of Analyzing Hazards and Controls where the group is ready for it, at a reduced pilot rate. Includes participant materials, evaluation results for your group, and the 30-day follow-up.

Pilots are delivered in person. CAWS is based in Maryland and covers the Mid-Atlantic without travel costs. Travel elsewhere is scoped into the engagement.

Pilot engagements are conducted under a written agreement providing for a structured debrief and for CAWS to use anonymized aggregate learning in openly published work. Individual responses remain confidential and are reported only in aggregate.

Ask about a pilot
Get in touch

Tell us who needs the training.

Say a little about the workforce, the systems already running, and roughly how many people. A short conversation is usually enough to tell whether a course fits.

hello@centerforaiworkersafety.org

Or use the inquiry form on the home page.

CAWS wants to hear from workers directly. If you are a worker and not here on behalf of an organization, write anyway.