Cam Stevens says artificial intelligence has huge potential in health and safety provided H&S practitioners accept the challenge of implementing it responsibly.

Author: Cam Stevens. First published in Safeguard magazine and reproduced with permission.

Key points:

• AI can shift H&S from prevention to prediction but needs careful change management to introduce successfully.
• Generative models and conversational assistants can help people make sense of complex H&S documentation.
• Computer vision with AI is already in widespread use to identify workplace risks and identify design improvements.

At the intersection of technology and workplace safety, Artificial Intelligence (AI) stands as a beacon of innovation, with the potential to redefine H&S management. AI’s role in enhancing workplace safety – shifting the narrative from prevention to prediction and enabling safer work design – is both critical and promising. Understanding how to responsibly leverage the potential of AI for H&S is arguably one of the biggest challenges facing our profession this decade.

Acknowledging AI’s expanding role in the workplace is the first step toward harnessing its potential. For H&S professionals, acquiring digital literacy is foundational, enabling meaningful engagement with AI’s workplace applications. Free resources like IBM and Microsoft’s introductory content on AI – and, perhaps ironically, AI-enabled search platforms like Perplexity, Google Gemini, and ChatGPT – provide essential insights, equipping individuals to navigate key AI use cases more effectively.

At the time of writing in early March 2024 – an important timestamp as the change of pace is rapid – the application of AI in health and safety is diverse, with numerous promising avenues being explored. There is significant experimentation with natural language processing, advanced analytics, and AI-enhancement of existing software solutions. Despite the wide range of use cases, two key applications of AI technologies have emerged as the most common.

Application 1: Generative AI

The first is Generative AI, a branch of AI that generates human-like and deeply contextual responses to user questions in real time. After being trained on sufficient learning material, AI systems can generate content in different formats, including text, images, video and even programming code, all from simple text queries. Conversational AI assistants, often referred to as copilots, provide a conversational interface with these generative AI systems.

A major benefit of these tools is helping H&S leaders to streamline safety administration, allowing them to spend more time in the field. Generative AI systems can rapidly summarise key data from lengthy documents, work instructions, email conversations, and more. With the right training of generative AI models, large volumes of static H&S procedures can be transformed into interactive and engaging formats.

One of the best examples of conversational assistants and generative AI being applied for H&S is the AISA (AI Safety Assistant) solution developed by P&O Maritime Logistics. According to Benjamin Neal, head of HSSEQ with the company. “AISA is an agile resource that can respond to employee queries anywhere in the world and in any language,” he told me. “Its real time functionality allows us to update company procedures or training material in one region and the data will be made accessible in all languages instantly.”

Application 2: Computer vision

The second and arguably the most well established application of AI for H&S is computer vision. This allows computers to understand and interpret visual information from images and videos, similar to how humans see and understand the world. It helps computers recognise objects, people, and actions in images or videos and to make decisions based on what they see.

H&S applications for computer vision are varied, from purpose-built hardware and software systems that provide real time alerts for high-risk plant and worker interactions, to automating ergonomic risk assessments using a smartphone camera.

These aren’t just theoretical use cases. Infrastructure construction company McConnell Dowell has recently showcased its use of computer vision technology, called Blindsight, developed by Australian tech startup Presien. McConnell Dowell uses Blindsight on mobile construction equipment to support the management of people and plant interaction.

Endeavour Energy, the operator of the electrical distribution network for a large area of New South Wales, utilises a computer vision solution from Soter Analytics to promote musculoskeletal health.

Change management essential

The deployment of AI solutions, particularly computer vision technologies, presents a compelling case for a comprehensive change management strategy.

Historically met with scepticism, the strategic introduction of these technologies requires transparency, communication and education to foster acceptance and trust.

This approach moves beyond mere surveillance, offering insights into workplace safety dynamics through advanced, real time analytics. Bede Cammock-Elliott, founder of New Zealand health & safety AI company seeo.ai, emphasises the importance of change management in deploying tech solutions.

“Customers who are enjoying incredible results are those who implemented the solution well,” he told me.

“They use it to confirm their problem hypothesis, remain flexible in the face of new or unexpected insights, have highly engaged leadership, closely partner with employees, share and talk about the number of events and their salience, communicate clearly, proactively (and constantly).

“Above all, they avoid the temptation to use the tech as a ‘stick’. One of the biggest learnings is to use the insights generated to focus on changing the work – work design – and not just attempt to ‘change the worker’ (ie: behaviour).”

Workers in the loop

Ashleigh Armstrong, head of partnerships at Soter Analytics APAC, further highlights the importance of worker involvement in the process.

“The key to successfully rolling out safety tech, especially AI, is to keep workers in the loop,” she told me.

“It’s about their day-to-day and might change how things are done, so let’s make sure they’re part of the conversation, excited and informed about what’s happening. At the same time, it’s vital that any new tech, particularly AI, actually makes work safer and more efficient. We need to design these AI solutions with the big picture in mind, focusing on real results in safety and business performance. Bringing people along on this journey and aiming for tangible improvements is how we make sure technology genuinely serves us all.”

Responsible innovation

AI’s application in health and safety, particularly with biometric data and video footage, demands a responsible innovation approach. This approach must be informed by the latest regulatory standards and ethical frameworks to ensure compliance and societal alignment.

The integration of AI into H&S practices presents an opportunity to significantly improve workplace safety. However, realising this potential requires strategic experimentation, robust change management, and a commitment to responsible innovation.

As we navigate the deployment complexities of AI in safety, the role of health and safety professionals in championing ethical practices and regulatory compliance cannot be overstated. By embracing these challenges with integrity and vision, we can leverage AI to create safer, more inclusive workplaces, marking a new era in health and safety management.

Cam Stevens is a safety technologist and safety futurist.

To learn more about Taronga Ventures portfolio company Presien and its AI vision system for worker safety, Blindsight, please visit https://www.presien.com/.