Artificial intelligence and drones

One of the most important tasks of unmanned aerial vehicles (drones) available today is data collection. The practice of this has raised many ethical and safety issues in recent years, so, among other things, legal regulations such as e.g. the GDPR, which, in addition to ensuring the protection of people’s personal data, makes the way they are handled transparent.

It is important to mention that, thanks to these strict rules, the way in which different organizations in Europe collect and process personal data has changed significantly. However, the question arises as to what standards and protocols apply to data collection when they are collected with drones. Most commercial UAS operating companies are not interested in information about where we are going, what we are looking at, or what we are doing, but are more interested in where, for example, there are cracks in a wind turbine, a pipeline. The commercial and industrial use of drones can make many of today’s activities much more efficient and safer, which can ultimately generate added value for society, either directly or indirectly. At the same time, the social acceptance of drones is still a major obstacle in many countries, as many are frustrated by the mere uncertainty and lack of information that a flying object buzzing over them in an unusual voice is carrying out some industrial / commercial activity or just being observed.

Drones are capable of capturing large amounts of visual data that can help law enforcement track potential emergencies and maintain public safety. However, for the data protection reasons mentioned above, it is challenging to authorize these types of activities and carry out projects. In a roundtable discussion at the Amsterdam Drone Week (ADW), which took place virtually on the COVID-19 pandemic in December 2020, participants discussed how to ensure the protection of people’s identities through artificial intelligence while providing law enforcement agencies with the ability to detect crime and find people in need.

A joint Amsterdam and aperdoorne test by the Amsterdam AI team, as well as the universities cooperating with them and the Dutch Police, has shown that artificial intelligence can be used to obscure details about people, cars and places. In this way, the police are not identifying objects and people, they are just detecting them: they will not specifically see and know that Aunt Hilda is walking down the street, only that a man can be seen in the picture.

First and foremost, people need to understand what each technological innovation serves and what impact it has on the lives of communities. The most effective way is to communicate all this in an understandable and unambiguous way to people and back it up with practical examples. Participants in the roundtable discussion also stressed that this process is slow and it will take a long time for people to develop trust in drone technologies. That is why it is important for individual societies to study thoroughly how and where technology can integrate into their national culture.