Nyheter

Les siste nytt fra Proactima her.

Tag: AI

How to use AI in emergency response

Imagine the following scenario unfolding in the North Sea. At 2am on Sunday, the pressure in the subsea pipeline carrying gas from the processing plant in Norway to Europe suddenly starts to drop, and the control systems also experience unusual signals and irregularities. Can artificial intelligence (AI) be used to streamline handling and decision-making in such an emergency situation?

Let’s put it to rest right away; AI is here to stay. And it’s already being used at various levels of emergency response today. Artificial intelligence (AI) has enormous potential to improve emergency preparedness and response. By leveraging AI technology, businesses can better predict, detect, respond to and manage different types of crises, from natural disasters to man-made events.

But first, how can we ensure that AI is used in the most effective and responsible way?

– We humans are vulnerable to large amounts of information that need to be processed. This is where AI can help us sort and structure the amount of information. For example, AI can help decision-makers by providing real-time analysis and forecasts based on available data, in other words producing decision support This can be particularly useful in chaotic situations where quick decisions are crucial. But, and this is a big but, we still have to use our human critical thinking skills since AI does not critically question whether we are making the right choices, nor can it make the decisions for us,” says Hanne Cook, preparedness and security expert at Proactima.

Hanne Cook and colleague and innovation expert Surbhi Bansal recently took to the stage at this year’s Beredskapskonferanse to explain how AI can be used to streamline management and decision-making in emergency situations.

So how can we actually use AI in emergency situations?

– AI models can be used to identify patterns or anomalies in data. This can be useful for detecting unauthorized activity, such as sabotage or terrorist attacks, as well as for monitoring cyber security and detecting hacking attempts. In terms of surveillance, AI can be used to monitor critical infrastructure such as bridges, roads and buildings to detect structural weaknesses or damage. AI can also be used to automate repetitive tasks, such as sending out emergency alerts, sorting information and coordinating task forces. This frees up human resources to focus on more complex tasks,” explains Surbhi.

Is it safe to use AI to develop decision support in an emergency situation?

– While AI can help us improve our situational awareness through the production of decision support, the question is what the risk picture will be when introducing AI into the management of emergency situations. Introducing AI-based systems is like introducing even more uncertainty and risk into an already complex system. KI can’t make the assessment for you and data quality is crucial. In the context of emergency preparedness, such tasks are not without risk. AI can also lead to uncertainty, risk, bottlenecks and poor explanations. This can often mean that the business does not have a sufficient understanding of how to control and verify AI-produced output and to what extent decision-makers should trust them,” says Surbhi.

From left: Hanne Cook and Surbhi Bansal during the Preparedness Conference 24. May 2024.

A (not entirely unlikely) contingency scenario

Back to our scenario. Since we already know that critical infrastructure is considered a likely target based on the security situation in Norway, is this sabotage or a technical error?

— In our subsea pipeline scenario, AI can be used to detect unauthorized tampering with equipment, monitor vessel movements in nearby and high-security areas, and detect suspicious activities in the network. In surveillance and monitoring, AI-assisted drones are already being used to patrol sensitive areas. They also have great potential for collecting data and estimating the trajectory of gas dispersion if the pipeline starts to leak. Subsequently, AI can be used to remotely control the delivery of necessary equipment and tools to quickly repair the pipeline,” explains Surbhi, adding that AI can also be used for communication, speech recognition and translations in the coordination work with relevant stakeholders.

Here are some ways AI can be used in emergency situations:

Prediction and prevention:
- Forecasts and alerts
- Risk assessment
Monitoring and detection:
- Real-time monitoring
- Detection of deviations
Response and coordination:
- Automation of routine tasks
- Decision support:
Communication and information:
- Language recognition and translation
- Information dissemination
After the event:
- Analysis and learning
- Recovery

Don’t forget human critical thinking

Hanne and Surbhi explain that the challenges of AI are amplified when dealing with an emergency situation – as these situations tend to be infrequent, stressful, fast-paced and highly uncertain. In order for AI to increase efficiency and provide good decision support in an emergency situation, Hanne and Surbhi encourage you to have:

Understanding how the technology works
Understand how machines arrive at what they show us
Understand how human factors come into play in a decision-making process
Human in the loop – human factors knowledge about the use of AI models in emergency management – to avoid the pitfalls that arise when humans have to work together with machines in a critical situation

Proactima already uses AI support to streamline risk analysis through our Dmaze tool. We have already come a long way in using AI to streamline processes before an incident occurs; risk analyses, exercise planning, support for preparing contingency plans, conducting exercises/training, etc.

—

Do you want to know more about how we use AI in emergency preparedness work today, or do you want input on how your company should work with AI in emergency preparedness? Get in touch with us for a chat!