Airports are quickly adopting a range of new technologies, from autonomous vehicles to artificial intelligence (AI), to improve passenger experiences and drive efficiency. One of the most significant ways in which AI is being implemented in airports has to do with security.
Across the globe, airports have implemented a number of new safety measures in recent decades because of emerging threats. In some instances, these measures have created significant delays when it comes to moving through the airport. As a result, the customer experience has suffered. This means airports are feeling mounting pressure to streamline the process as much as possible without creating any lapses in security.
Many airports, as well as the governments that back them, have looked to AI as a means of accomplishing the difficult feat of relieving wait times while maintaining strict security standards.
For example, the United Kingdom government recently invested 1.8 million pounds into the development of a new AI system in airports across the country. The Transportation Security Administration in the United States has implemented computed tomography (CT) scanners that implement AI to identify threats in several major airports. In other parts of the world, facial recognition has been implemented at customs and immigration stations.
These technologies do seem to have significant promise when it comes to improving security while minimizing wait times. Here’s how:
Machine Learning and Its Potential for Improving Security

One form of AI that shows the most promise is known as machine learning. AI systems can become more “intelligent” as they receive more information. In terms of airport security, systems can become very good at identifying threats based on patterns and do so much more quickly than a human could. Machine learning has driven trust in AI-based systems a great deal in the past few years.
Industry experts believe that implementing machine learning in airports could help avoid the need to scan certain items separately, such as laptops and other large pieces of electronic equipment. Letting passengers leave these items in their carry-on luggage as they pass through security would eliminate a significant amount of the delay caused in security lines.
One system employing this technology has already been developed. Called the Evolv Edge system, it uses cameras, millimeter-wave technology, and facial recognition to detect threats while people move through a scanner. The system successfully ignores non-dangerous items, such as keys and belt buckles, while reliably identifying explosives, firearms, and other weapons and hazardous materials.
Up to 900 people can pass through the scanner in an hour, making it not only more reliable than a traditional x-ray scanner, but also much faster. Evolv Edge is already being used to screen employees at the Oakland International Airport. It is expected to be deployed at other international airports soon.
How AI Could Reinvent the Airport’s Approach to Security
AI has already made a significant impact on airport security through biometrics. In the coming years, this technology will likely become even more widespread. A recent report showed that more than three-quarters of airports had new biometrics programs in the works for the coming five years. While people mostly associated biometrics with face scans, fingerprints and retinal scans are also expected to grow in popularity because of their reliability.
Some researchers want to go even further with this technology and have broached the possibility of behavioral biometrics. For example, University of Manchester researchers have created a system that identifies individuals based on gait and walking patterns as they step across a pressure pad. Each person has a distinctive, singular walking pattern.
Another application of biometrics is already being tested. The iBrderCtrl project involves an AI program in which a virtual border guard asks standard questions to individuals in an immigration line. If the system believes that the passenger is lying because of facial expressions, the individual gets passed on to a human for further review.
Of course, there is the question of accuracy with such technology. Early implementation of iBrderCtrl had a success rate of 76 percent. The developers believe that tweaks will make it 85-percent accurate. However, this rate may still not be acceptable to some airports, at least not as a primary means of maintaining safety.

The Challenges to Implementing AI-Driven Airport Security
Scrutiny of these systems remains high, especially after some prior failures. In the United States, $160 million was spent on body scanners that were later found to let serious threats through during undercover testing. Because of this, the standard is very high for AI technologies.
However, there is another hurdle that airports need to surmount: the amount of information collected by these systems, especially when it comes to biometrics. A great deal of information security is necessary to safeguard the privacy of passengers. In other words, while there is a lot of potential for AI in airport security, there are also significant challenges to overcome.



