The session will take a practical approach focusing on what organisations need to consider from a privacy, security, and governance perspective, while also exploring how AI can be used to improve health and safety outcomes in practice.

It will examine what practitioners should be thinking about as AI adoption increases, where it can genuinely add value, and the pitfalls to avoid. It will also reinforce the importance of critical thinking, professional judgement, and accountability in an AI enabled environment.

Michael Moyes, Partner, Duncan Cotterill

Sanja Marin, Senior Associate, Duncan Cotterill

AI is moving into recruitment, performance management, monitoring and investigations faster than most organisations can govern it. These tools are built on data about how "typical" people communicate, behave and perform. That makes them a new and largely invisible source of psychosocial risk for the up to 1 in 5 workers who think differently.

This session looks at AI through a neuroinclusion lens including where AI can exclude or harm neurodivergent workers, where it can remove barriers, and what safety practitioners need to ask before a tool is switched on.

Nicola Knobel, Head of Safety, Privacy and Compliance, Whānau Āwhina Plunket

Facilitator: Brent Sutton, Founder, Learning Teams Inc

AI is already changing safety work.

Some people are saying it will replace safety professionals. Whether that happens or not, one thing is clear: the role is changing fast.

The practitioners who understand how to build, question and challenge AI will be in a very different position from those who simply wait for someone else to decide how it will be used.

Because AI is powerful, but it is not automatically trustworthy.

It can hallucinate. It can reinforce bias. It can remove context. It can make uncertainty sound certain. And it can turn a complex event into a neat, convincing story that misses what really mattered.

In this practical 3-hour workshop, you will build a Microsoft 365 AI agent to help review investigation reports.

You will explore what makes investigation reports useful, where they go wrong, and what an AI reviewer should actually look for.

Then you will work through the full development and deployment cycle:

Learn → Define → Design → Build → Test → Improve → Deploy → Monitor

You will create instructions, knowledge and guardrails, test the agent against realistic investigation scenarios, deliberately try to break it, and improve it based on what you learn.

No coding or AI expertise is required. Ideally bring your laptop and have access to Microsoft M365 or ChatGPT.

You will leave with a working prototype, a practical method for developing AI safely, and a much clearer understanding of where human judgment still matters.

If AI is going to influence the future of safety work, do you want to understand how it works — or wait until someone else builds it for you?

AI may not replace safety practitioners. But safety practitioners who understand AI may have a very different future from those who do not.

If you are having issues, check out the below for some simple troubleshooting techniques. If you’re still having problems after following these steps, please let us know via the live support below (if available) and we’ll try our best to help.

If you experience issues during a live broadcast, please let us know. Pop a message in the live help support below (if available)

On some browsers, you may need to press the play button in the middle of your screen in order to start and view the stream *note that if the stream is not active you will not be able to play or view

To view in full screen, simply press the box or ‘full screen’ button on the bottom corner of the stream window. To exit full screen mode, simply press the ‘esc’ button.

Note the full screen icon may appear on the bottom left or the bottom right

Where possible, we attempt to broadcast in HD. The data rate for HD may be too large for your internet connection, so if you’re having issues, you can click on the settings icon in the lower right of the video window and select a lower speed/size which may help.

It’s normal that sometimes you can experience ‘buffering’ or ‘stuttering’, however if you are experiencing severe interruptions, then please let us know via the live help tab (if available) and we can look at adjusting our data rates.

t may not be the right time for the live stream event to start, so please double check you have entered at the right time. We’d also recommend checking your internet connection is operating properly.

On some browsers, you may have to press the ‘play’ button to start the stream.