Artificial Intelligence: Beyond Privacy
Opinion Column by Ivan Vandermeersch, the Secretary General of the Belgian Association of Marketing (BAM) and Board Member of FEDMA.
The joyous entry of digital society into our privacy thinking has shifted up a gear as a result of IoT (the “Internet of Things”). So-called ‘smart objects’ or ‘smart technologies’ collect data and even send it to the internet. Everyday objects are connected and can exchange data. Quo vadis, Privacy? And quo vadis, data protection authority?
Going against the stream is pointless. Driverless cars are being tested. Home automation consists of various adaptive applications for heating and lighting, optimised through data captured by sensors, sent to the internet and processed with algorithms. Robots will soon be able to speak and interact with people, socialize… It will be possible to take care of the elderly in their own home instead of locking them away in nursing homes. With the help of citizens, we will be able to carry out all sorts of measurements which are of interest to society or individuals, such as the impact of pollutants, bodymetrics, the energy consumption of a house, …
This is often fuelled by our personal data, unstructured or structured, numerical, from traditional database documents in unstructured text, e-mails, videos, audio formats and created by applications and sensors… Data sources are subject to a fast and continuous flow of new information generating large amounts of data. In realtime. This is a way for objects to generate information about projects, activities and the entire way of life of their users. Privacy legislation is struggling to keep up with technology.
It has been a long time since personal data was only analysed in a conventional way, and ever more often, it is the object of algorithms which connect data beyond geographical and computer boundaries, leading to the creation of an enormous mass of personal data which makes individuals identified or identifiable.
All of this is already happening, and the citizens who are the source of the data are by no means aware of the quantity and the nature of the information. The challenge consists of combining this mass of undifferentiated data and making it operational. By combining data, a series of personal information is created through multiple servers. The relevance of a datum depends on the quantity of other data it may be connected to. Automatically.
As a result, computers operated by human beings (desktops, tablets, smartphones) will come to represent a minority on the internet and will increasingly have to make way for internet users who will virtually consist of semi-intelligent devices, so-called ’embedded’ systems. Self-programming computers will interact with individuals on the internet, communicate with other objects and take autonomous decisions on this basis.
This is what “Machine Learning” is all about. It is a method involving automated data analysis through algorithms and learning from data in an ‘iterative’ way, i.e. based on repetitive processes, enabling computers to come to insights without having to be programmed by human beings. This is done using artificial intelligence: devices, software reacting to data or impulses from their environment and taking autonomous decisions on that basis. There may be a positive impact on medicine. There may be a negative impact on warfare. Applications equipping missiles with artificial intelligence already exist. And what if man loses control?
The Gafa (Google, Apple, Facebook Amazon – and let’s not forget Microsoft) are pioneers in this field. They are now being caught up by a Chinese little brother called BATX (Baidu, Alibaba, Tencent, Xianomi). The European digital desert is in between.
Europe must therefore focus on a respectful digital economy, with human rights such as the right to privacy as a foundation, and have the guts to steer clear of the pitfalls of the American, Russian and Chinese digital economies ‘sucking the digital life’ out of individuals. Europe could be the playing field of tomorrow’s digital world with techniques which are closer to the real economy by not tearing down current business models.
More and more often, work will consist of asking the right questions to these machines and programming them with the right ethical decision logic.
Algorithms are data driven, and the scope of a privacy policy no longer exclusively consists of databases which are analysed, but increasingly of their underlying algorithms. These are the basis for data gathering and the transformation into personal data. Consequently, this also involves the processes which were used to come to automated decisions.
Privacy policies can no longer exclusively be in the hands of lawyers locking themselves up in an ivory tower with their periods and commas. Understanding the digital challenges is a must for the new data protection authority. Technological progress and a rapidly evolving media landscape are creating a new balance between policy and society, with individuals at its heart. Data has become digital and digital has become data, all over the world.
Artificial intelligence will have to be given a heart and a conscience. Apart from privacy, political leaders must also be mindful of machine ethics. Europe needs data protection authorities whose vigilance reaches beyond personal data. Or will we leave this to Putin and Trump as well?
