Trainers, safety staff, and workers in every kind of job are using AI to draft course materials, procedures, safety briefings, translations, and evaluations, and the drafts often look finished while carrying errors that only someone who knows the standard will catch. CAWS runs hands-on sessions on where those errors come from, what an organization can put in place to catch and prevent them, and how to tell which uses of AI carry the most risk.
The AI in the Workplace courses teach people to recognize the hazards AI brings into their own jobs. This session is for organizations whose people use AI to produce safety information that others will rely on.
The session covers where AI errors come from, including training data that predates a changed standard and a built-in pull toward agreeable, confident answers, and which controls catch them, from keeping AI out of some uses entirely to approved source material, known-answer testing, and a named reviewer who signs off before anything reaches workers.
In the table exercise, small groups check AI output with planted errors against the governing standard, rate each use green, yellow, or red, and work out which control would have caught each error.
CAWS ran this session for about 70 construction union trainers at the annual trainer conference of a national construction safety training and research organization, with an answer key citing OSHA's construction standards that the host's technical staff reviewed beforehand. Trainers rated most of the uses red, the student evaluation scenario started a long discussion about where a better prompt helps and where it doesn't, and several trainers asked for a prompt library they could use in their own classes.
These are observations from one session with one group of trainers, not research findings.
The session above, built around the standards and materials your people actually work with, for a staff meeting, an instructor development day, or a conference. From $2,500 per session. See all prices.
A small working group from a few trades or teams picks the tasks they actually use AI for, tests prompts against them, and builds a shared library your organization owns. CAWS designs and facilitates the process.
Help writing your organization's guidance on when staff can use AI to develop training and safety material, who reviews and signs off on it, and what stays with a person.
Prepares your own trainers to deliver AI in the Workplace to your workforce without CAWS in the room. See all courses.
CAWS is not affiliated with any vendor and does not promote, sell, or certify any AI product or tool.
A one-page guide for anyone who uses AI to draft safety material. It covers what a prompt needs before AI can draft something useful, including the role, the audience, the length, and the exact source pasted in with an instruction to use only that source, with a worked example from construction built on OSHA's excavation standards, 29 CFR 1926.651 and 1926.652. It also explains how tone words quietly change the content and why the first answer should be treated as a draft.
Free to share within your organization.
Say a little about your sector, how your people are using AI to produce training or safety material now, and roughly how many would attend. A short call is usually enough to see whether a session fits.
Or use the inquiry form on the home page.