Artificial intelligence in the workplace: what’s in store for OHS?

Most organisations are still very much in their infancy when it comes to widespread adoption and application of artificial intelligence (AI) in the workplace, according to Microsoft Australia.

However, there is a stronger adoption of AI and investment in tools and the mechanisms to improve the workplace in industries with a greater degree of safety risk (such as mining, resources and agriculture) and a greater need for scale (such as healthcare), said Lee Hickin, national technology officer for Microsoft Australia.

“Perhaps one of the biggest inhibitors to many organisations’ ability to adopt AI into their workplace comes from the islands of data that are typically present,” said Hickin, who will be speaking as part as an AIHS virtual event on advanced technologies and innovations in workplace safety, which will be held on Thursday 3 September 2020.

“Building, inventory, people, work and asset data are often split across silos and so incorporating AI into the system first requires the consolidation of data into a single source.”

Hickin observed that many organisations are now embracing the potential of AI as well as building a culture that includes responsible and ethical AI at the core of their transformation.

However, data silos are a significant challenge for many organisations in terms of getting to the real insights and value that AI can provide.

“AI needs access to data to learn and get better and the integrity of the data is key,” said Hickin.

“For example, teaching an AI system to recognise dangerous spillages in the workplace requires the system to be shown (taught) many pictures of spillages of different types and severities so it can learn the difference between a water spillage vs a corrosive chemical spill which has a much higher severity.”

There are a number of ways in which AI technology can help identify and alert organisations to OHS issues in the workplace, Hickin added.

In the main, these come down to the use of visual image identification capabilities which use live and real-time image data to provide an alert on issues such as hazardous materials exposed, spillages and broken or damaged environments, incorrect use of safety equipment and the number of people in a constrained space.

AI technology can also be used to detect non-visible hazards such as poisonous gas levels, or sound or pressure changes in equipment that indicate an impeding failure, alerting people to the risk and activating a process of remediation automatically.

Hickin also observed that there are also a number of key trends occurring across a range of industries – the first of which is an overall reduction in the cost of being able to run the scale of data required to predict accurate AI models.

“It’s getting cheaper and therefore more accessible to individual and organisations of all sizes,” he said.

“Secondly, the availability of data is growing every day and this access to new and better data is helping to create more reliable and trusted AI systems.

“Finally, I would speculate that we are going to see an acceleration of AI at the edge – the ability for rich and powerful AI models to be run in small, low cost and low power devices able to be deployed into hundreds of physical environments.”

For organisations and OHS professionals looking to explore the adoption of AI, Hickin said it is important to have a data strategy and build a plan for how to access all data across the business.

The cost of doing this and the range of tools available to you in the cloud today make this much easier and less impactful to core business systems than ever before, he explained.

“Begin to pilot, test and experiment with AI – the technology is low cost and the data is freely available so do many experiments and learn about how AI can help your business,” Hickin said.

“Finally, take the time to learn and develop a knowledge of responsible AI. Understand the potential but also the risks of implementation that might lead to undesirable outcomes.”

Hickin will be speaking as part of an AIHS virtual event on advanced technologies and innovations in workplace safety, which will be held on Thursday 3 September 2020 from 1-4pm. For more information visit the AIHS virtual event series website.