Airport weapon detection: Why we need both man and machine

We have fast become used to automation in our personal lives – Siri, Alexa, robotic vacuums cleaners. You name it, we’re happy to hand over tasks that take precious hours out of our day over to machines. We use it to supplement our lives, often so we can be liberated from repetitive tasks and can instead devote more time to doing meaningful, impactful work. We don’t let machines run our lives, we just know working with artificial intelligence (AI) can help us achieve a lot more.

This is no different from the way industry is turning to AI, deploying automation to complement human skills. It’s certainly why automation is playing such an important role in weapon detection, whether at aviation passenger checkpoints, critical infrastructure protection, prisons or customs, applications across all these areas are starting to benefit from this technology.


Too often, we hear fears that machines are going to take over from humans, but this just isn’t the case – and it certainly shouldn’t be when it comes to identifying suspicious items such as weapons. In fact, it’s in this area where we need humans and machines to working together with their combined skillsets in order to get the best out of each other and guarantee all threats are spotted, 100% of the time.

AI excels at executing on repetitive tasks, analysing large amounts of data whereas humans are better at dealing with ambiguous information, making judgements and of course, interacting with people.

Keeping this in mind, we’ve built our iCMORE software on the premise that humans and machines need to work together to achieve the best results. Rather than either doing it alone, the software is built to help human handlers – security operators, customs officers and other controlling authorities better identify dangerous, prohibited or contraband goods. In addition to not getting tired and being unbiased, this technology will especially be useful to support less experienced operators.


Currently, each and every scanned image at security checkpoints needs to be assessed by a member of staff, whereas in hold baggage screening, only suspicious images are sent to an operator for analysis.  So why not replicate this process at the checkpoint to improve operational efficiency?

We are optimistic intelligent algorithms could lead to an even more automated screening process at checkpoints as well. The European Civil Aviation Conference (ECAC) and other regulators are starting to look into automated prohibited item detection – specifically around guns and knives – with a view to defining a common test method. This is great news, as new regulations definitely need to come into effect for automation to start dictating which bags are safe and which are not.

A combination of automatic object recognition with risk-based screening could prove particularly beneficial. Backing-up AI-enhanced screening results with passenger-specific risk-based data, might give a clearance to baggage without the intervention of an operator.

We think human intervention will remain necessary for the next couple of years, but it’s good to see more steps being taken into positioning AI as an assist for enhanced security. It will certainly ease the burden off image analysts, and help them to be more productive and accurate in their decisions. Ideally, we want a world where operators don’t have to spend time going through each and every image but, instead, can focus on analysing the ones that raise alarm.

Ultimately, weapon detection needs the strongest combination of skillsets to be successful. And, even with the most advanced AI technologies, this will always require a combination of man working with machine. Together, we can enhance detection process, through the positive attributes of both parties. In working as one, a system based on man and machine can be far more consistent as well as infinitely more thorough.


To learn more about how artificial intelligence can be used for automatic weapon detection at airports to increase security efficiency and detection accuracy click here.