Now 30 years in the manufacturing sector, I’ve watched the lightning speed of manufacturing technology advancement, and a growing divide as companies have struggled to keep up.

Canada’s tech adoption gap

In digital and automated technologies, manufacturers have access to undeniable new tools to reduce risk while increasing productivity. However, as of 2022, only two per cent of the industry had adopted robotics in their facilities, according to Statistics Canada’s Survey of Advanced Technology. The impact is clear, with the Bank of Canada noting the country has fallen to the second-least productive country in the G7 as of 2024.

Generative AI upside and risk

Today, generative AI and large language models are set to transform the industry yet again. Canada’s industrial sector has just begun to explore the potential for AI to streamline work, increasing efficiency and productivity. And while caution is common, most companies have little understanding of the potential new risks AI and cybersecurity threats represent to intellectual property (IP), employee morale, security, and physical safety.

Across industries, AI is a due-diligence issue. AI decisions affect confidentiality and IP, operational continuity, and regulatory exposure long before they show up on the shop floor. So, leaders who treat AI as they would any other high-risk operational change have a much better chance of achieving productivity gains without creating new safety and cyber risk.

Where do AI tools look for information, and how do they share it? As companies adopt these tools, leaders also must govern how they work.

Safety implications: Where AI helps and where it can harm

In safety, we need to consider how AI can best support our health and safety programs – and how it might derail them.

Training and guidelines help your employees leverage the potential benefit of AI, while using it efficiently and safely. Understanding the risks of AI to intellectual property, cybersecurity, and safety is essential knowledge for every business leader.

AI can summarize policies and data in seconds – a helpful aid to busy safety professionals. But it can also fabricate details or cite the wrong sources. For safety-critical work, a qualified human is required to verify every output.

Every day, generative AI models are learning, changing, and adapting. And we also need to learn and adapt to refine the results – to transform the AI models we develop with care into efficient and productive tools.

Workforce implications: skills and roles

When I started in manufacturing, robots were little more than a dream because they were so cost prohibitive for the average manufacturer. Today, companies are deploying used and new robots and cobots at much more affordable prices to automate operations. And with developments in 3D printing, manufacturers can now take ideas and concepts to working models and designs at a rapid pace.

Automation is eliminating tedious, repetitive, and dangerous tasks, moving workers into roles that require more examination, thinking, and intuition. Ultimately, automation is helping to shape safer roles for the humans in our production processes.

As automation and AI are helping to engineer out many manual labour and repetitive administrative tasks, they are also increasing demand for skilled labour. Heavy manufacturers are now looking for candidates with degrees in mechanical and electrical engineering and those with extensive technical skills. Others are upskilling and training employees to use new automated systems and perform routine maintenance as we make this historic shift.

This technology adoption is eliminating many unsafe jobs, creating new opportunities for upskilling, and attracting more highly skilled people to the industry. At the same time, it is introducing new areas of risk. To safeguard people and brand reputation, we are learning to look for risk in the intersection of people and robotics. We must consider new hazards in automated systems and the ever-growing risk associated with cybersecurity. And we need to address the very real risk of over-reliance on artificial intelligence that does not recognize the gaps in its own knowledge.

In the safety community, AI literacy and knowledge of automated systems is quickly becoming part of the essential skillset. Many of the same risks exist for practitioners: recognizing hazards in rapidly changing production environments; leveraging technology to keep people safe, while protecting these same people from technology errors; building critical thinking skills to leverage generative AI with healthy skepticism.

Next steps for leaders

The safest and most effective path forward begins with preparation. Define appropriate use cases for AI and, perhaps more notably, where its use is prohibited. Creating an AI usage policy is a critical step to protect the integrity of your IP.

Set clear expectations to require human validation for every safety-critical output. Update risk assessments as you implement new technologies. Document your cyber controls and, above all, train your people in AI literacy, confidentiality, and verifying AI-generated content.

The opportunity is huge. But proceed with eyes open, treating your implementation of AI as you roll out any new operational technology with a plan, effective controls, metrics, and a long-term plan for continuous improvement.

—

Editor’s Note:

This article by Alliance CEO Wayne Arondus was originally published in OHS Canada July 17, 2026.