AI in Airport Video Surveillance

According to data collected by the United Nations World Tourism Organization (UNWTO), an estimate of 1.45 billion people traveled in 2019, and it estimates that the number will rise to the north of 2 billion in 2020. With the number of travelers increasing, airports are under enormous pressure to ensure that passengers are processed and screened both securely and safely. Since the turn of the new millennium, due in part to the number of people choosing to travel and in response to the subsequent emerging threats, airports worldwide have significantly increased security and surveillance measures, primarily through the implementation of advanced video surveillance technology. AI-powered cameras and machines are operating as virtual security guards in airports around the world. In fact, in the opinion of the multinational information technology company, SITA, almost 80% of the air transport industry is planning to implement artificial intelligence technology by 2021. It is expected that this will improve security, which could mean a more secure traveling experience for the traveler.

Artificial intelligence is defined as the intelligence demonstrated by machines enabling them to perform human-like tasks, such as facial and speech recognition, decision-making, and the translation of languages. Research into AI is conducted to create technology that allows machines to function effectively and intelligently. Advancements in artificial intelligence are accomplished through the use of training datasets. Datasets are a collection of data. In the case of AI, datasets are used to teach the machine to recognize things: this works by providing the system with a set of labeled photographs and videos. For example, if researchers want to train it to differentiate different breeds of dogs, they would need a large inventory of images of all the different dog breeds with labels. Similarly, for it to recognize normal behavior and distinguish it from abnormal behavior, pictures would have to be provided with both labels and context. AI systems improve as more information is added into them; by combining large amounts of data with intelligent algorithms, the machine can learn from patterns it observes and revamps automatically through experience.

AI-powered analytics are training computers to monitor surveillance cameras. Here’s a look at how artificial intelligence is being utilized in airport video surveillance:

• They can detect when luggage has abandoned and can identify by whom.
• The facial identification feature serves to identify individuals who are on a no-fly list.
• They are capable of recognizing when people remain within specific areas for longer than a defined time.
• The use of cameras as a virtual tripwire triggers an alarm when a passenger enters an unauthorized zone.
• The system detects any camera tampering attempts, and this includes covering, redirecting, or spraying the device.
• They can send alerts to airport authorities when a person is moving in “the wrong direction” within a designated area.

The Future of Artificial Intelligence

The future of AI- analytics is bright. Researchers are currently working on not just training computers to monitor human behavior but also to comprehend human life and interactions. Here are what AI-camera systems have in store for aviation security.

• Passenger Identification: AI technology has made the process of passenger identification quick and easy for airport officials through the use of self-service check-in systems and facial recognition, which is currently being used to scan travelers as they go through customs. Advancements in AI technology could eventually lead to a system through which travelers are identified without having to present a boarding pass, as fingerprinting and retinal scans are expected to become increasingly popular for security purposes at airports.
• Detecting Suspicious Behavior: Although it is relatively cheap to gather video surveillance, it remains expensive to monitor and analyze that data. At the moment, humans are responsible for detecting potential threatening behavior in passengers, which leaves a lot of room for error because their performance is influenced by many factors, such as mood and emotions. But with artificial intelligence, the detection of threats will be consistent and reliable and quicker as technology like this can process large quantities of people; this is because the machines will be able to provide a thorough screening of passengers by collecting their data from predictive analytics, behavior tracking, and pattern recognition and by checking their booking and travel history. And even from data gathered from their social media accounts.
• Detect Dangerous Items: AI-powered technology can automate the process of baggage and passenger screening. By utilizing such techniques, belongings that currently need to be scanned separately, such as cell phones and laptops, will be left in passenger luggage while they move through security checkpoints. They can also detect weapons hidden in baggage and beneath clothing as well. Machine learning techniques can be used to detect metallic and non-metallic weapons such as firearms, knives, and all types of explosives, including plastic devices, on a person; and this is possible through the AI-powered computed tomography (CT) scanners’ ability to produce 3-Dimensional images of an individual’s body, then analyze the photos to search for any items that don’t belong on it. Cameras equipped with this type of technology can be positioned outside of airport terminals so that threats can be detected before they have even entered the airport.
• Maintenance Prediction: Airports are exploiting the technological advancements being made by launching several initiatives aimed at boosting the reliability of aircraft maintenance. They have plans to implement AI technology to anticipate potential problems with aircraft maintenance, through its ability to collect and store data. The technology’s predictive maintenance analytics will make aircraft maintenance easier by alerting the crew on how often and when it should be done.
• Improve Customer Service: AI is being used to assist customers in the airport. Airlines companies are also using AI systems to help their customers gain access to answers to frequently asked questions, as well as provide accurate information on the status of a flight. These systems not only benefit the customer, but they are also simultaneously reducing the operational and labor costs of airline companies.

Consequences of Artificial Intelligence in Cameras

Here are five problems that AI surveillance will cause:

• Uneasiness: The rise of AI-powered surveillance will lead to an environment where people are always self-conscious and anxious, wondering what constitutes as normal behavior and adjusting their actions as not to trigger alarms, out of fear of being stopped by a security guard or placed on a watch list; this will be a consequence of the cameras that will be continuously scrutinizing every move made and continually making judgments as to whether the behavior is normal and whether the authorities should be notified.
• Privacy Invasion: Any behavior or physical trait that a machine learns to recognize it can also remember and report; this means that every surveillance camera will be a digital checkpoint, that utilizes facial recognition to capture people’s identities, associations, and locations. And this type of surveillance possesses a considerable threat to a person’s privacy.
• Deceptive AI Cameras: One primary concern over the use of these technologies is their accuracy; AI cameras could potentially incorporate analytics that is untested or deceptive. With the demand for advanced surveillance systems being at an all-time high, there is a possibility that some systems being sold will not work as advertised. Analytics that are inaccurate, if authorities give them credence, can lead to passengers being falsely identified as a threat. And even if the security guard clears them, the incident could be entered into the person’s record and stored by the machine and used against them later on. Even tested and authenticated AI cameras have shown unintentional and hidden biases within the systems. While some algorithms do not incorporate race or gender explicitly, they do include variables that have a distinct racial or gender bias, such as clothing or hairstyles; such prejudices could lead to embarrassment when middle eastern passengers are misidentified as “suspicious,” even though they weren’t on the authority’s radar.
• Abuse of Power: Data collected by AI cameras could fall into the wrong hands. As robust surveillance infrastructures become more readily available, they’re creating opportunities for misuse at the hands of whoever controls them. Who knows what the consequences of this level of data collection and privacy invasion could be!
• Violation of the United States Constitution: Some users of AI analytics could be in violation of the United States Constitution, especially as they become more and more thorough and intrusive. The collection of physiological responses such as heart rate and pupil dilatation could be deemed a violation of the Fourth Amendment’s law against unreasonable searches of an individual and their belongings.

Recommended Solutions

As the aviation industry continues to embrace the advantages of artificial intelligence, protective actions need to be taken to prevent the grim consequences of AI video analytics. The data collected by AI-powered surveillance systems should be regulated by practices that ensure reliable privacy protection. As this field grows, policymakers should focus on creating laws that strictly govern the sharing, retention and destruction of information. Information that is not deleted but retained, however, must apply audio and video redacting software tools to conceal the passenger’s identity.

The redaction of audio and video recordings protect the privacy of both the passengers and airport staff. Automated redaction software utilizes artificial intelligence to obscure sensitive information, such as faces, voices and travel documents. Laws that restrict access to the types of data collected from government and private sector agencies should also be put in place; this prohibition from accessing and retaining (or even purchasing) would be a mechanism to prevent abuse of private information, as well as the collection of passenger data for marketing purposes. Additionally, due to the fact that AI-powered video analytics have shown discriminatory behaviors, they should be annually subjected to independent testing and review by expert researchers to identify unintentional biases, ensuring algorithmic fairness and transparency.

Conclusion

Although artificial intelligence-powered surveillance technology such as behavior and face recognition and body scanners may be unsettling to some, they are needed in the effort to prevent potential life-threatening incidents and ensure passenger and staff safety. With the application of artificial intelligence and their capabilities, the aviation industry will eventually be able to strengthen its security to impenetrable levels, enhancing its protection without compromising the experience and enjoyment of the traveler. But to avoid a privacy violation catastrophe, it must also invest in measures such as video and audio redaction software that will reduce the harmful consequences of artificial intelligence.