Artificial Intelligence (AI) has steadily made its way into various industries, one of the most important among which is of infrastructure relating to public safety. As compared to the human workforce and its efficiency, strategically programmed software and unmanned aerial vehicles such as drones can contribute to public safety more efficiently and at a much faster pace. Here, we focus on the uses of AI and drones in overseeing and enhancing safety, and a brief foray into matters of privacy and malfunctioning related to the same.
Drones have come to occupy a place of great importance in the fields of security, medicine, law enforcement, and natural disaster relief. They are now used by both government and private enterprises to supervise, assess, surveil and in many cases transfer supplies to places that are hard to access on the ground.
Natural disasters like floods, earthquakes, and cyclones are much better mitigated by an aerial survey that helps in locating and assessing areas prone to and worst hit by a disaster. For instance, Hurricane Harvey in Texas and Hurricane Irma in Florida saw massive use of drones in spotting locations hit by the storms, as well as spotting victims stuck in inhospitable locations. Drones play a vital role in airlifting medical supplies to rural and inaccessible places, getting external defibrillators faster than ambulances. Today, the World Health Organization is experimenting with Matternet’s quadcopters to reach supplies to Bhutan.
One of the most important applications of drones is in fighting fires. Drones are able to drop fire retardants as proximally as possible, capture close images of the color and velocity of the fire, and survey post-fire situations to ensure that every person has been evacuated. All of this not only reduces life risks to firefighters but also cuts down on costs by systematic planning.
Like drones, Artificial Intelligence (AI) can be skillfully used in preventing, lessening and mitigating crises. The most important advantage of AI is that it can monitor and analyze human behavioral patterns to prepare databases which can then be used to increase security.
It can thus be used in the prevention of crimes at local and national levels by generating a large database of criminal activities and monitoring recruitment to terrorist organizations. In the unfortunate event of a crime already committed, AI can help in speedy investigation and recovery of illegal weaponry. South Africa, for example, has combated wildlife poaching at Kruger National Park and gun violence in Cape Town using ShotSpotter technology. Terrorism can thus be kept in check by monitoring of propagandist literature on social media. Such data collection can be done by AI faster and with greater precision.
Additionally, international borders are often secured by planting AI checkers at crucial places. These are especially important today, considering the worldwide travel restrictions due to COVID-19. Recently a new AI-powered system that can detect coughing sounds has been developed, with the intent of understanding how and where the disease is spreading. Developments like these could be especially useful at international airports.
At airports, current AI-systems can gather, analyze and inspect passenger information, the content of their luggage and their travel plans. The most prominent example of this is the AI lie detector called iBorderctrl, employed by the EU at various airports. Non-EU travelers must get their passports and visas scanned, answer a series of questions by an AI border guard, and address a webcam that scans and analyzes facial movements to detect lies. The accuracy of this AI system is expected to rise to 85% in the near future.
The Potential for Errors
It goes without saying that the application of such technological facilitators must strive to be error-proof. Organizations depend heavily on the use of drones in cases of fires, natural disasters and reaching of medical supplies. Such a vital technological emergency service provider cannot afford to malfunction. However, accidents do happen – consider the Swiss drone crash of 2019, where a drone crashed merely 50 yards away from a group of children, resulting in the cessation of drone delivery. For drones to be effective in public safety, malfunctions like this need to be eliminated.
Likewise, a lot can be improved in AI in terms of software relating to facial recognition and recording of human behavior. The inherent bias present in AI applications could easily result in increased racial profiling, risking the rights of many individuals. There is also a very important concern of privacy — AI is often developed around the unethical and often unannounced use of large amounts of personal data. In order to gain the optimum use of AI, governments and private enterprises must ensure ethical and legal transparency and involve some amount of human intervention in decisions made by AI.