BELGIUM

Automating Society 2019

Introduction EUROPEAN UNION RECOMMENDATIONS ABOUT US

Belgium Denmark Finland France Germany Italy Netherlands Poland Slovenia Spain Sweden United Kingdom

By Rosamunde van Brakel

Belgium has several levels of government and debates are spread out over the different governments: federal, regional (Flemish, Walloon and Brussels-Capital region government) and the community governments (Flemish, French and German). In general, the current Belgian political discourse uses the terms ‘Big Data’ and ‘Artificial Intelligence’ (AI). In general, the governmental strategies concerning digitisation and AI are embedded in an economic discourse and are about increasing jobs and supporting companies in the context of Industry 4.0. Although automation and digitisation has been going on for a long time, in Belgium this has always been characterised by problems with implementation. The result is that Belgium is rather behind the rest of Europe, especially when it comes to digitisation of the public sector.

Political debates on aspects of automation – Government and Parliament

Digital Belgium: Digital Agenda – Federal Government

The federal government does not have a specific strategy concerning AI or ADM, but they did launch “Digital Belgium: Digital Agenda” in 2015. According to this agenda, by 2020 it should be possible for Belgium to get into the top three of the European Digital Economy and Society Index, for 1,000 new start-ups to take root in Belgium, and for the digital revolution to deliver 50,000 new jobs in a variety of sectors.

“Digital Belgium“ is a plan that outlines the long-term digital vision for the country and translates this into five priorities of the federal government:

Digital infrastructure
Digital confidence and digital security
Digital government
Digital economy
Digital skills and jobs [BE 1]

Recently, Belgian prime minister Charles Michel commented on AI, saying that he “is convinced that we will need new professions in the future and that they will be made possible by this technological evolution”, and that he is “convinced that artificial intelligence is an opportunity for quality of life, to advance the quality of medicine, telecommunications, and to raise the standard of living on this earth.” [BE 2]

The Belgian Privacy Commission published 33 recommendations about Big Data in 2017. Most of the recommendations refer to the GDPR, especially when it comes to automated or semi-automated decision-making. [BE 3]

Radically digital Flanders – Flemish Government (Region and Community)

The Flemish government launched its digital strategy: Vlaanderen radicaal digitaal (Radically digital Flanders). It is inspired by the Digital Agenda of the Federal government and has the same priorities. [BE 4]

The Flemish government also has a specific programme on Artificial Intelligence that looks at how AI can improve government services. Five programme directions are identified as:

The human computer: digital = super handy, ‘thinks and integrates’ like people thanks to chatbots and conversational platforms
The computer assistant: hyperpersonalisation in government services by collecting data about citizens—and starting from that knowledge—offer a more personal government experience.
The super-quick (proactive) computer: a quick smart government as a result of text, language and image recognition and other pattern recognitions.
The autonomous computer: more with less. Through automation of tasks, civil servants can do other work.
The moral computer: digital ethics. We are actively following European initiatives. [BE 5]

The Flemish Minister of Innovation Philippe Muyters announced that the government will invest €30 million in AI. In his “AI action plan”, he presents three main goals: fundamental research, applications for industry, and framing policy concerning education, sensitisation and ethics. PwC, an international consultancy, has been appointed to do an international benchmarking exercise of Flanders to get a better picture of where Flanders is at concerning AI. According to Muyters, the “potential societal and economic impact of AI is enormous. For Flanders, the biggest opportunities lie in the first instance in personalised healthcare, smart mobility and industry 4.0. If we tackle this evolution quickly and smartly, we can make sure Flanders will reap all the benefits.” [BE 6]

The Flemish government states that research programmes that are proven to be of international excellence will be strengthened and deepened, that the government will make “clear choices on the basis of excellence, so that budgets will not be fragmented but instead will be invested in areas where there is the highest potential. Special attention will go to leading AI-technology platforms with clear market potential.”

The Innovation Ministry believes that Flanders “can be one of the frontrunners for the application of AI in the business community. This can be done, not by inventing everything, but rather by functioning as a living lab for Flemish and international applications.” So-called priority clusters and Vlaio (the Flemish Bureau for Innovation and Entrepreneurship) are supposed to “take care of knowledge sharing and to establish a network to follow AI trends and translate these to Flemish companies.”

The Flanders government states that there is a need for a broad sensitisation to the disruptive potential of AI technology: “Both in education and in the corporate world people are working at installing permanent training provisions. In addition, an AI think tank will be established to examine the ethical implications that AI entails.”

Several events have taken place in 2018 under the direction of the minister. In July a “stakeholders forum on Artificial Intelligence” was organised. [BE 7] In September, a conference and exhibition took place to show the potential of AI: SuperNova [BE 8].

A parliamentary question in the Flemish Parliament on October 3, 2018 discussed the above-mentioned plan of minister Muyters. High on the Flemish political agenda is the need to develop new degrees at universities and training in Artificial Intelligence [BE 9].
A new masters degree in Artificial Intelligence has already been launched at the Katholieke Universiteit Leuven [BE 10].

Digital Plan – Walloon Government

In December 2015, the Walloon government adopted its digital Plan du Numérique (“digital plan”) [BE 11]. Its main goal is to become a major Industry 4.0 player and a forerunner in the digital revolution. It is inspired by the Digital Agenda of the Federal Government and has the same priorities.

Political debates on aspects of automation – Civil Society and Academia

Ligue des droit de l’homme

The Ligue des droit de l’homme is a Walloon non-profit organisation for human rights in Belgium. They have a commission looking into the consequences of new technology for human rights. The commission initiates actions and activities that allow it to get in touch with and to react to the population and / or to motivate the creation of citizen initiatives. To this end, the commission is responsible, alone or in collaboration with other actors, for setting up activities or projects. The commission is also responsible for examining files, drafting working papers, articles and position papers, setting up or intervening in conferences or other awareness-raising activities, initiating or participating in action plans, bringing challenges to the courts and confronting the public authorities on the themes within its competence. [BE 12]

Liga voor Mensenrechten

The Liga voor Mensenrechten protects human rights by denouncing structural and incidental violations to create a societal foundation for human rights in Belgium. They do this by informing, taking action, and going to court. They work on privacy and other human rights issues related to new technologies, and they organise the Big Brother Awards. They do not work specifically on ADM or Artificial Intelligence, but they are increasingly looking at these technologies especially in the context of the Big Brother Awards. [BE 13]

Privacy Salon

Privacy Salon is a non-profit organisation, which aims at sensitising and critically informing the broader public, policy makers and industry in Belgium, Europe, and beyond about privacy, data protection and other social and ethical issues that are raised with the introduction of new technologies in society. Privacy Salon organises the annual CPDP conference where several panels focus on ADM. One of the main themes that is being worked on by the organisation is algorithmic discrimination and algorithmic decision-making. More specifically they are organising events on this theme including an art exhibition and a workshop on Algorithms and Society. [BE 14]

Royal Flemish Academy of Belgium for Science
and the Arts (KVAB)

The KVAB published an opinion piece in 2017 on Artificial Intelligence. The main purpose of this document was “to inform the public as objectively as possible and to propose a series of conclusions and recommendations to concerned parties in order to deal with AI and ensure that our community can properly benefit from its huge opportunities, as well as get an insight into the risks and what to do about them.” Also, The Class of Natural Sciences (KNW) of the KVAB has started a working group to study the impact of AI in Flanders. [BE 15]

Regulatory and self-regulatory Measures

Belgian Law concerning the protection of data of natural persons in relation to the processing of personal data

The most important law regulating automated decision-making is the data protection regulation. The Belgian version of the GDPR, and the replacement of the 1992 Belgian Privacy Law [BE 16], came into force on September 5, 2018. The law applies to every fully or partially automated processing of personal data, and also to the processing of personal data which is not automated, but which is included in a file or will be included in a file. [BE 17]

‘Killer robots’

The Belgian Chamber of Representatives adopted a resolution in 2018 to have a preventative ban on fully automated weapons (‘killer robots’). [BE 18]

ADM in Action

Algorithmic work activation

The public employment service of Flanders, VDAB, together with the Katholieke Universiteit Leuven and Vlerick Business school developed algorithms that provide insight into the way people search for jobs on their website. [BE 19] The system analyses thousands of job seeker files and looks at the click behaviour of people who are looking for jobs on the VDAB website. According to the VDAB, this process has an important predictive value concerning long-term unemployment. The information is supposed to allow for early and more efficient intervention. One goal for the VDAB is to see if click behaviour analysis can be used to control the active search behaviour of the job seeker. The job seeker who is not active enough online would then be invited for an interview and the next step would be a penalty [BE 20]. Another application would be similar to Amazon’s recommendation system. On the basis of the huge amounts of data VDAB collects, it could then provide the person with a list of recommended jobs and present potential employers to the right candidates. According to the VDAB, by using these data-driven methods it is possible to improve personal guidance of jobseekers. [BE 21]

Predictive policing

In 2016, a local police zone on the Belgian coast started implementing predictive policing software. The chief commissioner claims that since the start of the project, criminality has gone down by 40 %. According to the police, the types of criminality that the predictions are the most effective at are burglaries and vehicle theft as there is a lot of data available about these crimes that can be analysed by the software. Via the data that the police receives, they claim that they can predict in which neighbourhoods it is more likely that a burglary will take place. On the basis of this prediction they will send out an intervention team. The chief commissioner wants to expand the system by interconnecting the software with Automatic Number Plate Recognition (ANPR) cameras. [BE 22] [BE 23]

In 2016 the Belgian federal government invested in the iPolice system to centralise all police data in the cloud. The system should be operational by 2020. This is a cooperation between the Ministry of Home Affairs, the Digital Agenda and the Ministry of Justice. In an answer to a parliamentary question, the minister of Home Affairs, Jambon, stated on October 27, 2016: “The new technologies should make possible a better linkage, sharing and analysis of information in a quick way. The police should work and act on the basis of an integral analysis of structured and unstructured data, from internal and external available data.” [BE 24]

In September 2018, Federal and local police issued a press release to say that they have big plans for predictive policing and already see the possibility that, from the next legislature (after the council elections of October 14, 2018), predictive policing experiments can begin in Antwerp and other local police zones. According to the spokesperson of the Federal Police, they are still working on the tools and building the systems. The data that will be used for the analyses will come from the police databases, for instance the frequency that certain crimes appear in certain areas. In addition, data from external sources will also be important. Predictive policing is mostly seen as a tool to help the police do their work more efficiently. [BE 25]