Can artificial intelligence (AI) be used in fire safety?

Artificial intelligence (AI) is an aspect of technology that holds both many advantages and pitfalls for the fire-safety community and general public, but it’s one we can’t ignore. So, what can we expect from its use in fire safety?

Fire investigation

I have for many years found that investigating serious or fatal fires is very difficult if you are not part of the investigation team or a relevant authority and this often means that when you examine any official report it generally raises many questions that the authorities are sometimes reluctant to answer or clarify. To try to answer these questions it requires a Freedom of Information (FOI) request to the relevant authority, and whilst some authorities are quite helpful, others can be very guarded in their responses, and this leads me to ask why should this be?

Could it be that the relevant authority are protecting their own reputation and are not willing to divulge any information that may show where they have not acted in a manner that was in the interests of the general public or other interested parties, or could it be that they are simply covering up mistakes that they made in dealing with the incident or enforcement?

I have always thought that the enforcing authorities should be more open, honest and transparent in these matters as it is only by an approach like this that we can truly learn from these incidents, and this must be in the interests of everyone and not just the enforcing authorities.

It is interesting to note that in some European countries the investigation is done by the police who also have forensic science teams available to them.

Artificial intelligence (AI)

One of the first questions that I asked AI was ‘Can AI be used for fire safety and investigation?’ And this was the response:

‘AI can be used in fire investigation in a number of ways. For example, AI can be used to:

• Identify the origin and cause of fires. AI can analyse data from the fire scene, such as burn patterns, heat distribution and the presence of accelerants, to identify the point of origin and the most likely cause of the fire.
• Reconstruct the sequence of events leading up to a fire. AI can use data from the fire scene, as well as witness testimony and other evidence, to reconstruct the sequence of events leading up to the fire. This can help investigators to identify suspects and determine whether the fire was accidental or intentional.
• Predict the behaviour of fires. AI can be used to simulate the behaviour of fires under different conditions, such as wind speed, humidity and fuel type. This information can be used to help firefighters develop effective fire suppression and evacuation plans.
• Identify fire hazards. AI can be used to analyse data from building inspections and other sources to identify potential fire hazards. This information can be used to develop fire-safety regulations and to help businesses and homeowners to reduce their risk of fire.
• AI is still under development, but it has the potential to revolutionise fire investigation. By automating tasks that are currently time-consuming and labour-intensive, AI can help investigators to solve fires more quickly and efficiently. Additionally, AI can provide investigators with new insights and perspectives on fire behaviour and fire investigation.

‘Here are some specific examples of how AI is being used in fire investigation today:

• Fire investigators are using AI to analyse burn patterns. AI can be used to identify patterns that would be difficult or impossible for human investigators to see. This can help investigators to identify the point of origin of a fire and to determine the most likely cause.
• AI is being used to reconstruct the sequence of events leading up to a fire. AI can be used to analyse data from the fire scene, such as the location of fire debris and the presence of heat indicators, to reconstruct the sequence of events leading up to the fire. This can help investigators to identify suspects and to determine whether the fire was accidental or intentional.
• AI is being used to predict the behaviour of fires. AI can be used to simulate the behaviour of fires under different conditions, such as wind speed, humidity and fuel type. This information can be used to help firefighters develop effective fire suppression and evacuation plans.
• As AI continues to develop, we can expect to see even more innovative and effective ways to use AI in fire investigation.’

So, clearly there is an important role for AI in Fire Investigation but unless all of the relevant information is made available, I wonder how accurate the results will be and whether they will be made available to interested parties.

Bibby Stockholm barge – is it safe for the intended use?

This is a question that I was recently asked by the BBC and as all the authorities and manufacturers were not being very open about the answer, I decided to ask AI the question and this was the response, but please note that the accuracy of all these responses is not guaranteed:

‘Fire safety is of paramount importance on board the Bibby Stockholm barge. The barge is equipped with a number of fire-safety features, including:

• A fire detection and alarm system
• Fire extinguishers
• Fire hoses
• Fire blankets
• A fire escape plan

The crew of the Bibby Stockholm barge are trained in fire-safety procedures and are responsible for maintaining the fire-safety equipment on board. In the event of a fire, the crew will activate the fire-alarm system and extinguish the fire using the fire extinguishers or fire hoses. If the fire is too large for the crew to extinguish, they will evacuate the barge and call the fire brigade.

‘Here are some additional fire safety tips for the Bibby Stockholm barge:

• Keep fire exits clear and unobstructed.
• Do not smoke or use open flames on board.
• Store flammable materials in designated areas.
• Inspect fire-safety equipment regularly and make sure it is in working order.
• Hold fire drills regularly to ensure that the crew is familiar with the fire safety procedures.
• By following these fire safety tips, the crew of the Bibby Stockholm barge can help to prevent fires and ensure the safety of everyone on board.’

I then asked AI: ‘Is the Bibby Stockholm barge safe for 500 people?’ and this was the response: ‘Whether the Bibby Stockholm barge is safe for 500 people is a matter of debate. The barge was originally designed to accommodate 222 people, but its capacity has been doubled by putting bunk beds in its cabins and converting some communal rooms into dormitories. This has led to concerns about overcrowding and fire safety.

‘The UK Fire Brigades Union has warned that the barge is a ‘deathtrap’ and that the government is putting people’s lives at risk by using it to house asylum seekers. The union has also expressed concerns about the lack of training and experience of the staff who will be working on the barge.

‘The Home Office has defended its decision to use the Bibby Stockholm barge, saying that it has been thoroughly inspected and is safe to accommodate 500 people. The Home Office has also said that the barge will be staffed by experienced and qualified staff.

‘Ultimately, whether the Bibby Stockholm barge is safe for 500 people is a matter of opinion. There are valid concerns about overcrowding and fire safety, but the Home Office has insisted that the barge is safe.

It is important to note that the Bibby Stockholm barge is currently being used to house asylum seekers. Asylum seekers are particularly vulnerable people who are often fleeing persecution and violence. It is therefore essential that the Home Office takes all necessary steps to ensure that the Bibby Stockholm barge is a safe and humane place to live.’

Clearly, in this example AI is not able to fully answer the question but does provide a good overview.

The Royal Clarence Hotel fire

This is an incident that I know a little more about as I carried out an investigation for the BBC and a subsequent programme was shown on the BBC Inside Out programme.

The results of my investigation were published here https://www.ifsecglobal.com/global/frs-report-royal-clarence-hotel-fire-full-holes-missing/ and following this I was contacted by one of the fire officers that attended the fire and who indicated that my investigation was nearer to the truth than the official report. This led to two programmes on the BBC that produced a detailed investigation into the firefighting operations, together with eyewitness reports that also contradicted the official version.

The AI version gave the following response:

‘Dave Jones, a former sector commander of Devon and Somerset Fire and Rescue Service (DSFRS), believes that the hotel could have been saved by the fire service. He claims that he told incident commanders that he had found a fire on the third floor, but that his evidence to an inquiry was left out of a fire service report.

‘Jones says that he needed authority to send crews back into the hotel, but that he was met with blank stares. He also claims that 90% of the hotel was perfectly intact, and that it could have and should have been saved.

‘An independent review into the fire concluded that it remained speculative as to whether the historic hotel could have been saved. However, the review did make a number of recommendations for areas of learning for future incidents dealt with by Devon & Somerset Fire and Rescue Service.

‘Devon and Somerset Fire and Rescue Service has denied any wrongdoing and says that it hopes the review will reassure the public. However, Jones’ claims have raised serious questions about the fire service’s handling of the incident.’

I would add that I also received a number of other responses from serving members of DSFRS that confirmed Dave Jones version of the story but were unwilling to be named due to possible action by their employers.

How can we develop AI to be more helpful in fire safety?

This is what AI tells us:

‘There are a number of ways in which we can develop AI to be more helpful in fire safety. Here are a few examples:

‘Improved fire detection: AI can be used to develop more sophisticated fire-detection systems that can detect fires earlier and more accurately than traditional systems. This can be done by using AI to analyse data from a variety of sources, such as smoke detectors, heat sensors and cameras.

‘Predictive fire prevention: AI can be used to develop predictive models that can identify potential fire hazards and predict where and when fires are likely to occur. This information can then be used to prevent fires from happening in the first place.

‘Firefighting assistance: AI can be used to develop systems that can assist firefighters in fighting fires. This can be done by providing firefighters with real-time information about the fire, such as its location, size and intensity. AI can also be used to develop robots that can help firefighters to extinguish fires and rescue people.

‘Here are some specific examples of AI-powered fire-safety technologies that are currently being developed:

• AI-powered smoke detectors: these detectors can use AI to identify different types of smoke, including smoke from different types of fires. This can help to reduce false alarms and improve the accuracy of fire detection.
• AI-powered fire surveillance cameras: these cameras can use AI to detect fires and smoke in real time. They can also be used to track the movement of fire and smoke, which can help firefighters to better understand how a fire is behaving.
• AI-powered fire prediction models: these models can use data from a variety of sources, such as weather data, building data and historical fire data, to predict where and when fires are likely to occur. This information can be used to prevent fires from happening in the first place, or to deploy resources more effectively in the event of a fire.
• AI-powered firefighting robots: these robots can be used to extinguish fires, rescue people, and explore dangerous environments. They can also be used to collect data about fires, which can help firefighters to better understand how fires behave and how to fight them more effectively.

‘These are just a few examples of how AI can be used to improve fire safety. As AI technology continues to develop, we can expect to see even more innovative and effective fire-safety solutions emerge.

‘In addition to developing new AI technologies, we also need to make sure that existing AI technologies are used safely and effectively in fire-safety applications. This means developing and implementing best practices for the development, deployment and use of AI-powered fire-safety systems. It also means educating firefighters and other fire-safety professionals on how to use AI-powered fire-safety systems safely and effectively.

‘By developing and using AI responsibly, we can make a significant contribution to improving fire safety and reducing the loss of life and property due to fires.’

Can AI be used to carry out fire risk assessments?

This was another question that I asked AI, and this was the response:

‘AI can be used to carry out a Fire Risk Assessment (FRA). FRA is a process of identifying and evaluating the fire risks in a building or structure, and then developing a plan to mitigate those risks. AI can be used to automate many of the tasks involved in FRA, such as:

• Identifying potential fire hazards: AI can be used to analyse building plans and other data to identify potential fire hazards, such as flammable materials, electrical wiring and heating systems.
• Evaluating the risk of fire: AI can be used to evaluate the risk of fire based on the identified hazards, as well as other factors such as the occupancy of the building and the fire-safety measures in place.
• Developing a fire-risk mitigation plan: AI can be used to develop a fire-risk mitigation plan that includes measures to reduce the likelihood of fire, as well as measures to protect people and property in the event of a fire.
• AI-powered FRA tools can help businesses and organisations to save time and money, and to improve the safety of their buildings and occupants. However, it is important to note that AI-powered FRA tools are not a replacement for human expertise. It is still important to have a qualified fire-safety professional review the results of any AI-powered FRA and to provide guidance on the implementation of the fire-risk mitigation plan.

‘Here are some of the benefits of using AI for FRA:

• Accuracy: AI can be trained on large datasets of fire-safety data to identify fire hazards and assess fire risks with a high degree of accuracy.
• Efficiency: AI can automate many of the tasks involved in FRA, which can save businesses and organisations time and money.
• Objectivity: AI is not susceptible to human biases, which can help to ensure that FRA is conducted in a fair and objective manner.

‘However, there are also some challenges to using AI for FRA:

Data quality: AI-powered FRA tools are only as good as the data they are trained on. It is important to ensure that the data used to train the AI is accurate and up to date.

Transparency: it can be difficult to understand how AI-powered FRA tools make decisions. This can make it difficult to trust the results of the assessment.

Accountability: if a fire occurs, it is important to be able to determine who is responsible for the failure of the FRA. This can be difficult to do if the FRA was conducted by an AI.

Overall, AI has the potential to be a valuable tool for FRA. However, it is important to be aware of the challenges involved in using AI for FRA and to take steps to mitigate those risks.’

Conclusions

I have spent nearly 60 years in the fire and safety profession and without doubt this is one of the most exciting developments that I have seen in my career, but as with all new technology it will need collaboration by everyone in the industry and especially fire and rescue services who need to be more open, honest and transparent in their actions.

I have also been discussing AI with my colleagues in Europe and they are also seeing this as a great step forward.

As can be seen from the potential development areas there is hardly an aspect of fire safety where AI cannot be a useful tool, but we do need to be careful that it doesn’t take over because at present we cannot guarantee the results.