AI and Safety Management: Smarter Safety Without Losing the Human Touch

Show Notes

What happens when artificial intelligence meets real-world safety management? In this episode, Sheldon breaks down how AI can support Environmental Health and Safety professionals in practical, meaningful ways. From hazard recognition and data analysis to training development, trend spotting, and management decision-making, this conversation explores where AI fits into modern safety systems and where human judgment still matters most.

This episode is for safety consultants, EHS managers, trainers, and leaders who want to understand how to use AI as a tool for better planning, better communication, and better outcomes. Sheldon cuts through the hype and focuses on what matters: how AI can help improve safety performance, support compliance efforts, and make safety management more proactive instead of reactive.

If you’ve been wondering whether AI is just a buzzword or a real advantage for safety professionals, this episode gives you a practical lens for using technology without losing the people-first foundation of EHS.

What is The Safety Consultant Podcast?

The Safety Consultant Podcast with Sheldon Primus is your ultimate weekly guide to starting or growing a profitable occupational safety and health consulting business. Are you ready to be your own boss and make a greater impact? Your expertise can help more people create safer workplaces, and your skills deserve a platform where they can truly shine. If you feel limited in your current role and believe your knowledge could serve the broader workforce more effectively, this podcast is for you. Join us as we explore the steps to launch your own safety consulting business, share insights on navigating the industry, and provide strategies to maximize your impact on workplace health and safety. Now is the perfect time to take control of your career and make a difference!

[SPEAKER_00]: Welcome to the Safety Consultant Show with Sheldon primus, where we blend safety sparks with business brilliance and the pitch of Sheldon's signature wit, whether you're a safety pro of fresh face consultant, or just safety curious, get ready for a show that will educate, entertain and elevate your safety game.

[SPEAKER_00]: Let's dive into the art of consulting with your safety shirt for himself, Sheldon primus.

[SPEAKER_01]: This episode is powered by Safety FM.

[SPEAKER_02]: Welcome to this show, everybody.

[SPEAKER_02]: This is the show that I teach you the business of becoming a safety consultant.

[SPEAKER_02]: Look about OSHA compliance, we get some really good guests.

[SPEAKER_02]: So, welcome to the safety consultant podcast.

[SPEAKER_02]: Today, what I really wanna go over and have you guys here is my fault on AI and the use of AI in our industry.

[SPEAKER_02]: Now, we're all professionals out there, so we're starting to integrate AI a little bit more.

[SPEAKER_02]: No, it's just going to touch a little bit about how I do it, how I teach my students on how to integrate that.

[SPEAKER_02]: So today's conversation is going to be on the use of AI.

[SPEAKER_02]: So when we come back, give it a try.

[SPEAKER_02]: Tell you guys how I eat.

[SPEAKER_02]: You get into the safety business because you want to protect people, but somewhere along the way the child started controlling your time, your schedule, your priorities, even your income potential.

[SPEAKER_02]: What if the expertise you've built over the years can the field give you something more?

[SPEAKER_02]: More freedom, more control, more ownership of your future.

[SPEAKER_02]: I'm Sheldon Prime as the author of seven steps to starting a profitable safety consulting business and the host of the Safety Consultant podcast.

[SPEAKER_02]: But more than 350 episodes helping safety professionals become consultants and grow their career.

[SPEAKER_02]: Consulting changed my life.

[SPEAKER_02]: In fact, I'm a full-time RVer and a one consulting contract paid for my home I live in today.

[SPEAKER_02]: Now, I help safe professionals build their own consulting practice through biweekly coaching calls, the accountability and personal guidance.

[SPEAKER_02]: But this program is for everyone.

[SPEAKER_02]: Start with the consultant readiness audit at shelvin.coach.

[SPEAKER_00]: We now return to the safety consultant show.

[SPEAKER_00]: Once again, here's Sheldon primates.

[SPEAKER_02]: All right, so, well, let's get into it.

[SPEAKER_02]: So, not everybody's going to be using the generative pre-trained transformers.

[SPEAKER_02]: There's a bunch of them out there.

[SPEAKER_02]: I use ChatGVT mostly.

[SPEAKER_02]: I've been hearing good things about Gemini and there's Claude.

[SPEAKER_02]: There's a bunch out there.

[SPEAKER_02]: So, co-pilot, that's the one Google uses.

[SPEAKER_02]: There's just so many out there.

[SPEAKER_02]: But what it's truly doing is just honestly predicting what's the next thing to go, you know, in a sequence if you would.

[SPEAKER_02]: So how do you work this thing?

[SPEAKER_02]: What do you do with it?

[SPEAKER_02]: Now let you guys know that I try to train my GPT and I've got a couple of them and I say couple because I have started using my own, I kind of created my own GPT.

[SPEAKER_02]: So, chat is the one I've been using the most, but honestly I started training a chat to have a safety consultant and then one for learning management.

[SPEAKER_02]: But overall, I've been trying to...

[SPEAKER_02]: get more confidence in it, right?

[SPEAKER_02]: And that's the key.

[SPEAKER_02]: What you're always going to hear people say is you just can't put in a prompt and then let it go.

[SPEAKER_02]: Just listen to whatever the GPT spits out and make that your final work.

[SPEAKER_02]: Safety in health, we can't do that especially because if we're dealing with concentration or chemicals or anything like that and we're trying to use the AI assistant to help us troubleshoot or to help us try to figure out thinking out of the box type stuff.

[SPEAKER_02]: Then we got to be really careful about what we do and how much we simulate this information.

[SPEAKER_02]: I've seen then heard of people just truly using AI to create their written programs and not a good plan, unless you know what you're looking for.

[SPEAKER_02]: And that's one of the key things that I always say is, with the AI, you're going to need to know yourself something about that topic.

[SPEAKER_02]: So, for instance, if you're going to use AI to create a presentation.

[SPEAKER_02]: All right, that's something very easy.

[SPEAKER_02]: You could do a format at RealWell for you.

[SPEAKER_02]: Depends on what system you use.

[SPEAKER_02]: We'll have different capabilities with graphics and all that stuff.

[SPEAKER_02]: That's a good use for it.

[SPEAKER_02]: You pop in your information that you're looking for.

[SPEAKER_02]: And then now you get a presentation out of that, right?

[SPEAKER_02]: The thought process is that if you're going to use it for advanced things such as here is a case study that we have from an incident and I'm looking for a lesson learn then at that point if you're trying to figure out a lesson learn and you're going to create some sort of recommendations only using what AI spits back at you.

[SPEAKER_02]: You can be opening up yourself for either some information that is just not relevant or it's information that is just not something that you should be using as a professional.

[SPEAKER_02]: So what you'll need to do is you'll need to know about the topic.

[SPEAKER_02]: You'll need to know about the case itself.

[SPEAKER_02]: You'll need to know context.

[SPEAKER_02]: so that when the GPT spits something out at you and whatever it comes back as you're now going to look at it as an assistant talking to you and you have to verify what the assistant says and you're still gonna be spending some time on this product not just regurgitating what this thing just gave you and just do a copy and paste for a mirror your AI assistant

[SPEAKER_02]: So, I've kept many, many times, where the AI system, I would ask it, and I would do demonstrations in my class, and I would ask that to give us, like for instance, I would ask to give it with the class, some sort of incentive program, and I would give it the actual parameter of what I'm

[SPEAKER_02]: what my budget is and all that stuff and then I'll let the program work and inevitably I would have something come back as one of the criteria for this program to work this safety safety incentive program is how many days without an injury or illness and you do that in real life you're getting yourself a notion of fine if you're in the US and any regulatory

[SPEAKER_02]: It's not going to like that matrix because what's that saying is how many days you did not report an injury, you're going to get some sort of reward injury may happen, but your reward will come from not reporting that injury.

[SPEAKER_02]: air illness and that's when you get yourself in trouble.

[SPEAKER_02]: So I, you know, you got to know the context first and then you have to adjust it.

[SPEAKER_02]: So I'm not saying don't use it.

[SPEAKER_02]: I'm actually telling people to embrace the use of AI.

[SPEAKER_02]: We're in my mind, I think, where the first generation, and I believe that there's going to be more and more generations that are going to end up using AI and even more of just technical and smart ways.

[SPEAKER_02]: Now, we're now seeing AI be used in so many applications that we don't even know.

[SPEAKER_02]: So I believe it's time for us as professionals to embrace it.

[SPEAKER_02]: And what I would tell my students and I'll tell you guys now in this episode is put in your query, whatever you want the AI to figure out, but you're not going to just take it blindly.

[SPEAKER_02]: Your time isn't going to be spent in formatting and some of the thinking your time is going to be spent in going into research and figuring out the research part of it and truly see

[SPEAKER_02]: Does this thing have or is it truly what it says it's supposed to be so for instance?

[SPEAKER_02]: You're doing something in Ocean Standard and now I've had this happen, so that's why I know this.

[SPEAKER_02]: I literally had an AI's response add to section or to say sub-part E in 1910 which is going to be the general industry standard.

[SPEAKER_02]: Sub-part E is your exit routes and all that.

[SPEAKER_02]: Emergency action plans and fire prevention plan.

[SPEAKER_02]: It added an additional section that did not, that there was no section in the standard at all.

[SPEAKER_02]: And literally read out or reply to me in an actual training, it gave an extra training standard, and it gave it a number and everything in 1910, section 40 something.

[SPEAKER_02]: I knew it wasn't there, so if you weren't familiar with that standard, you would just cite it and then at that point you're going to lose credibility as well as you're going to make it where your organization is going to lose.

[SPEAKER_02]: again, credibility, but lose faith in the safety department, but then now the workers aren't going to trust you, and I get a trust what you're saying, you're fooling it in at some point, and if you didn't know to look for that, then it could have gone past your desk to someone else's desk.

[SPEAKER_02]: So what you need to do is first you get to know a little bit about the topic before you start sharing this information.

[SPEAKER_02]: So that's the key and you're going to spend your time well, don't let the formatting and all that stuff related to how the final thing looks.

[SPEAKER_02]: You're going to spend your time on the verifying content verifying that this is truly what the standard saying or whatever the problem is.

[SPEAKER_02]: So that's the key to the AI.

[SPEAKER_02]: It's going to save you some time.

[SPEAKER_02]: But the time it's saving you is going to be more formatting.

[SPEAKER_02]: And out of the box thinking, and then your time is going to be sent, spent verifying the information, and then seeing how viable everything is, and you're going to go back and forth with it.

[SPEAKER_02]: It's supposed to be a system to you, so you keep asking questions, if you've been feeding it information, such as a full CFR code of federal regulation,

[SPEAKER_02]: standard and keep feeding and information to help it be smarter, help it predict the next thing better.

[SPEAKER_02]: So, your use with the AI is going to be where it's part of the team, but it isn't the decision maker you are.

[SPEAKER_02]: You're going to be the one who gives the final approval your name is on there, your credibility is on the line, so it's not the AI.

[SPEAKER_02]: You're going to definitely have to work with that.

[SPEAKER_02]: I had on my own situation where I used AI for creating some of my blog posts on shellandprimus.com and we could find some blog posts on there.

[SPEAKER_02]: I'm still going through some of the things that was posted on that blog post and I did it through an AI assistant that, honestly, you just gave it some keywords

[SPEAKER_02]: I didn't do that many, but I did a bunch, so the idea was to drop people to your site using the keywords.

[SPEAKER_02]: So I noticed after, well, I actually, I didn't notice first, but my daughter saw some stuff, and she's a virtual assistant for me.

[SPEAKER_02]: And she's like, well Dad, you can't use AI like that.

[SPEAKER_02]: ended up having to retract some things and hurt my credibility so from there I stopped using that program and honestly just started thinking about what can I do to make sure I don't lose credibility anymore.

[SPEAKER_02]: So again,

[SPEAKER_02]: AI is good.

[SPEAKER_02]: You want to start using it.

[SPEAKER_02]: You want to start getting familiar with how to do prompts for a familiar following up on each answer familiar on feeding your AI assistance really good information to train from and then you're going to go ahead and

[SPEAKER_02]: truly take that information, make it your own.

[SPEAKER_02]: And then in some cases, I also do a disclaimer if I'm using a significant amount of this information.

[SPEAKER_02]: I'll do a disclaimer, especially if I'm using formatting.

[SPEAKER_02]: And I'll have a disclaimer on my final document that there was AI use in the development of this product.

[SPEAKER_02]: And some of the critical thinking is mine, but the formatting is AI, so some combination of that.

[SPEAKER_02]: So what I'm looking for you guys to do is...

[SPEAKER_02]: I can't familiar with your AI or whatever you're using, but then I'll see you on again.

[SPEAKER_02]: Keep an eye on the final product.

[SPEAKER_02]: Be willing to look through everything, read it all, and then make your changes.

[SPEAKER_02]: is not going to be the final document, so just make sure you know that and you be aware of that.

[SPEAKER_02]: But again, this technology is not going away, we want to be responsible for it, we can use it for critical thinking, we can use it for trying to figure out the predict what could be the next.

[SPEAKER_02]: How's this thing that's going to happen?

[SPEAKER_02]: Or best practice?

[SPEAKER_02]: The excellent use.

[SPEAKER_02]: So now you're going to take your expertise and you put it into your assistant.

[SPEAKER_02]: That's where it's for.

[SPEAKER_02]: And then now at that point you get yourself a good final product, right?

[SPEAKER_02]: So just let me know in the brace to future.

[SPEAKER_02]: Just be cautious.

[SPEAKER_02]: There you go.

[SPEAKER_02]: That's the message.

[SPEAKER_02]: So...

[SPEAKER_02]: Pick your, your whatever service that she is, and then just go ahead and give it a try.

[SPEAKER_02]: Don't be scared.

[SPEAKER_02]: You got this.

[SPEAKER_02]: Go get him.

[SPEAKER_00]: That's a wrap for today's safety show.

[SPEAKER_00]: Remember, safety isn't just about hard hats and caution signs.

[SPEAKER_00]: It's about business smarts too.

[SPEAKER_00]: Thanks for joining us on the safety consultant show with Sheldon primus.

[SPEAKER_00]: Until next time, stay safe, stay savvy, and keep consulting like a boss.

[SPEAKER_00]: Go get him.

[SPEAKER_00]: The views in the opinions expressed on this podcast or broadcast are those of the host in its guest and not necessarily reflect the official policy or position of the company.

[SPEAKER_00]: Examples of analysis discussed within the past hour only examples.

[SPEAKER_00]: It should not be utilized in the real world as the only solution available, as they are based only on very limited and dated open source information.

[SPEAKER_00]: Assumption made within this analysis are not reflective of the positions of the company.

[SPEAKER_00]: No part of this podcast of broadcast may be reproduced stored in a retrieval system or transmitted in any form or by any means mechanical electronic recording or otherwise without prior written permission of the creator of the podcast of broadcast Sheldon Primus.