Kaylie Moncur, Head of Marketing at SeeChange, sheds light on the transformative power of artificial intelligence in fire safety

For hundreds of thousands of years, humans have used and controlled fire.

From cooking food to powering industrial revolutions, fire has been a tool that humans have both used and respected.

With this respect for the danger that fire represents, humans have developed regulations for using it safely.

In the early days of humanity this could mean not lighting a fire around dry foliage, humans having witnessed the devastation of wildfires on the landscape; for modern life there are potential fire risks at every turn from retail storage to petrol filling stations.

So how do we keep pace with safety when the risks are constantly evolving?

OSHA and HSE are regularly updating standards in line with new risks, but for businesses and individuals these standards are both numerous and rely on human interaction with the standards to maintain compliance.

A fire safety officer can write a detailed fire risk assessment, plan safe routes out of the building and comply with the legislation on paper.

However, there will always be unknown additional factors that emerge, and the human nature to take shortcuts to overcome.

Despite the regulations and implementations of risk management and fire prevention, fire and explosions (excluding wildfires) are the single largest identified cause of corporate insurance losses, representing 21% of all claims with more than €18 billion of insurance claims between 2017 and 2021.

Human nature is to assume risks are lower, especially in familiar environments, and confirmation bias can reinforce this – an employee left a box of products in front of an electrical panel a week ago and it was fine, it will be fine this time too.

The employee may be right, this time it might be fine, but what about next time?

What about the day of an unannounced fire safety inspection, or the day that the electrical panel becomes faulty and sparks?

Businesses can be meticulous planning fire safety on paper, but in the end they rely on employees to stay within the rules.

When this falls down, this can lead to fines as a minimum.

Fires can cost in building damage, insurance premium increases, compensation for anyone caught in the fire, and in worst case scenarios it can cost lives.

What does AI have to do with Fire Safety?

Artificial Intelligence (AI) is (along with connectivity, human-machine interaction and advanced engineering) one of the four foundational types of what is referred to as the fourth industrial revolution.

Computer Vision AI seeks to understand scenarios by extracting meaningful information from the surroundings, and to execute actions defined by programmed rules.

Introducing AI into Fire Safety allows businesses to balance human actions with agreed safety parameters and to demonstrate to health and safety regulators, insurance companies, their own employees, and customers that they take fire safety seriously.

Although it can seem complicated to introduce AI to a business environment, solutions can be designed to run on edge devices, integrate with existing CCTV, and be fine-tuned to individual environments, providing 24/7 cover.

 A challenge for Health & Safety professionals is to apply regulations consistently across sites where layout differs.

AI can enforce rules designed by fire professionals, allowing them to focus on reducing or eliminating associated risks.

An example could be the electrical panel scenario we previously highlighted.

Whilst health and safety professionals can decree access must not be blocked with a minimum clearance of 30″x36, how do they enforce this when employees are running the store day to day?

Without a health and safety professional on site throughout store opening hours, it is reliant on employees remembering and actively following risk assessment guidelines.

For businesses with plenty of employees, this can be less challenging, but for those with fewer employees trying to complete more work, health and safety may not be a priority.

By turning on AI solutions which monitor for scenarios such as these, businesses translate the risks outlined by their health and safety advisors into actionable rules, with the health and safety management system integrating with existing communication methods.

In store notifications would then be sent advising items have been stored incorrectly and requiring them to be relocated.

Of course, AI like any tool has its limitations.

 For example, data requirements are a challenge – AI needs to learn about its environment, and this typically requires huge amounts of data.

 In addition, there is always a risk that if an AI system encounters something in an environment it is not familiar with, something important could be ignored (known as a False Negative) or something irrelevant could be flagged (a False Positive).

Overcoming limitations such as these requires a combination of machine learning and human oversight.

 By ensuring solutions are built with human inputs this enables AI to make safer and more accurate decisions, and for them to continually improve by taking these corrections into account.

 It also ensures that humans can review and understand the decisions AI makes and intervene with this decision-making process when necessary.

What can AI offer for Fire Safety?

The key areas of fire safety that AI can really help businesses succeed with are:

1. Continuous monitoring of protected zones including exit routes, exit doors, electrical panels, with rules and parameters determined by both the regulations and the individual business layout.
2. Early alerting for spills of flammable and hazardous liquids such as fuel leaks at a petrol station
3. Notifications for obstructed access or missing firefighting equipment such as fire extinguishers.
4. Insights into incidents, near misses and general store safety
5. Reporting and audit trail documentation to demonstrate to regulatory bodies a commitment to ensure compliance

For businesses this is an extension of their existing health and safety practices, enabling them to demonstrate to regulatory bodies that they understand and comply with regulations.

It also provides opportunities for employee training or business process improvement – is a fire door always blocked at 9am on a Monday because of stock deliveries?

Are there more obstructions in exit routes during peak holidays due to excess stock deliveries with no place to store it?

Businesses can utilise AI to guide their employees into complying with the regulations.

But in addition, when used properly AI can also provide a wealth of data including critical insights, incident reports and audit trails to help understand what is happening at a granular level across the business.

For retailers with multiple stores in a chain, deploying a health and safety management system can allow them to understand the entire retail estate immediately – which stores are more frequently non-compliant, which regions may need more employee training, which areas are performing well and so consistently strive for better standards.

This can then enable businesses to reward the stores performing well, and support those who are not compliant.

Stores for example with high staff turnover may not have enough employees with the awareness and training to enforce regulations.

Some stores may be understaffed and lack the capacity to get the store cleanliness and fire safety hygiene under control.

For businesses, this would allow them to assess the situation as a whole and address problem areas.

In the example above, if a store is understaffed, businesses may focus more attention on recruitment drives in specific regions, as this is likely to be much more cost effective than repeated fines for non-compliance.

Finally, AI can enable remote cross-estate fire safety compliance visibility.

AI allows managers to see how stores perform day to day, how fast employees respond to fire safety non-compliance and how many incidents there are every day, week, and month.

Businesses can then use this data to find ways to incentivise their employees to adhere to continual compliance – stores with less than 10 incidents per quarter may get a bonus or other benefits for maintaining health and safety compliance.

Whilst the world of AI seems vast and it can be difficult for businesses to know how or where to start, it is easy to see the value of AI in these settings.

Not only can businesses demonstrate how seriously they take fire safety to regulatory bodies, but they can ensure that workplaces are truly safe for their employees and customers.

This article was originally published in the April 2024 issue of International Fire & Safety Journal. To read your FREE digital copy, click here.