AI at Work: Transforming Health & Safety with Practical Intelligence

 Success with implementation of AI is just 20% tech and 70% is about people. You need to help your team overcome that initial barrier when they feel AI isn't useful. That's where the real gains are. 

- Dmitry Grigorenko

About this session

Artificial Intelligence is no longer a future concept; it's a tool reshaping health and safety today. But the real value lies beyond the hype.

In this session, we shift the focus from theory to hands-on practice. Craig Blakeley sits down with Dmitry Grigorenko, Regional WHS Manager at Amazon (UK) and author of AI for Health and Safety (2022), for a practical, no-nonsense look at how generative AI is actually being used by WHS professionals — from intelligent risk assessment and training tools to AI-driven assistants that support proactive decision-making.

Dmitry walks viewers through live tool demonstrations, the psychology of learning a new skill like AI, and the honest limitations — hallucinations, data confidentiality, and AI's total lack of logic — that every safety professional needs to understand before relying on it.

Avoiding hallucinations

Dmitry brings practical prompting habits to the table, asking AI to cite sources, rate its own confidence, and flag guesses — to catch fabricated facts and figures before they end up in your work. Plus why AI should never be trusted with maths. 

Key takeaways from the session

•  The Dunning-Kruger Effect, Applied to AI 
  Why early AI users often overestimate what they've achieved, hit a wall when results feel “generic,” and give up right before the point where real skill, and real value, begins. 
  
•  The Four Stages of Building AI Skill
  From basic prompting, to expanding your toolkit through practice and exploring others' use cases, to developing an intuitive feel for how AI responds, to finally scaling that skill across your whole team. 
  
•  Confidentiality and Data Safety 
  What actually happens to the data you share with AI, why the real risk sits in what leaves your servers rather than in the model itself, and practical steps for using AI safely without organisational tools in place. 
  
•  The Research: AI Adoption Is Still Early
  A 2023 BCG global study shows leaders were using AI far more than frontline employees, most organisations were still in early-stage experimentation, and successful AI adoption came down to 70% people, 20% technology.
•  Real-World Use Cases 
  From trend analysis across incident data and multilingual, context-aware translation, to generating CPD reflective statements for IOSH members and using AI as a coaching or mentoring “conversation partner.”