Public safety is a democratic responsibility for All. So is protecting the freedom to assemble. Europe must do both without any discrimination.
On 16 September 2026, researchers, law-enforcement practitioners, policymakers and civil-society representatives met at Vrije Universiteit Brussel for the GATHERINGS Final Conference, examining how surveillance at public gatherings can be made more effective, transparent, privacy-friendly and socially responsible. The EU-funded project, coordinated by VUB, set out to develop common standards and a practical framework for balancing security against privacy, economic costs and social impact. [1]
The subject is politically urgent. Demonstrations, festivals, sporting events and public celebrations are essential features of democratic life, but they can also present genuine security challenges. Authorities may have to respond to violence, theft, vandalism, threats or disorder, as we have seen lately in different manifestations. In some circumstances, cameras and other surveillance technologies can assist situational awareness and investigations.
Yet the existence of a security risk does not make every surveillance measure necessary, effective or proportionate. Nor does technological sophistication guarantee better public protection and nudges avoidance.
For young European federalists, this is not simply a technical debate about cameras and algorithms. It is a question about democratic power, sovereignty: how should public authorities protect people while ensuring that the exercise of that power remains lawful, accountable and open to scrutiny?
The answer should not be a choice between security and freedom. It should be a European model of effective security under democratic control.
- Start with evidence, not fear
Public debate about demonstrations is often shaped by striking images: clashes with police, damaged property, looting or large crowds in tense situations. These events deserve serious attention. But exceptional incidents cannot, by themselves, tell us how common such events are, what causes them, or which interventions prevent them.
Belgium has a public source for examining recorded crime: the Federal Police’s crime-statistics portal. It provides downloadable data, reports and geographical breakdowns, including information by region, municipality and police zone. Its categories cover recorded offences and specified criminal phenomenon. [2]
This makes it possible to investigate patterns in theft, property damage and other offences in Brussels and most of the European cities. It does not, however, mean that the published figures automatically identify offences committed during demonstrations. To establish that connection, researchers would need event-specific police records, official incident reports or other documented evidence.
That distinction matters. A rise in recorded property damage across a municipality cannot automatically be attributed to demonstrations, just as a fall in general crime cannot be attributed to CCTV without evidence linking the intervention to the outcome.
A responsible security debate should therefore ask four empirical questions:
- How frequently do serious offences occur at different kinds of public gatherings?
- What threats were authorities responding to in each case?
- Which interventions were used, and what evidence shows whether they helped?
- What were the financial, privacy and social costs?
These questions are more useful than assuming that every gathering is dangerous or that installing additional cameras will necessarily make it safer.
Therefore is relevant to highlight that our first democratic obligation is to describe the problem accurately. The second is to demonstrate that the proposed response works.
- What do public surveillance actually achieve?
CCTV can have legitimate uses. Depending on the context, it may help monitor an area, support a response to an incident, provide evidence after an offence or assist an investigation. But those potential benefits must be distinguished from claims that cameras prevent all forms of violence or disorder.
The Campbell Collaboration’s systematic review of CCTV and crime examined controlled studies comparing areas with and without camera interventions. It found that effects varied according to context and crime type. Its review is a useful reminder that the effectiveness of surveillance should be assessed empirically rather than assumed. The research does not establish that cameras prevent looting at demonstrations in Brussels specifically. [3] Also, Rob Teslemans highlighted the fact that with his experience in large-scale camera project in several cities in Belgium, people’s behavior could be opposite and some tend to reduce their public presence in an area where there is more cameras ( due to privacy or they assume the place is more prone of a crime to appear).
The difference between detecting an offence and preventing it is particularly important. A camera might help identify a suspect after property has been damaged without having prevented that damage in the first place. A recording might also be inconclusive, inaccessible or insufficient to identify the person responsible.
Likewise, the number of cameras installed is an input, not an outcome. A large network does not, on its own, establish how much crime was prevented, how many investigations were assisted, or whether the system offered value for money.
The relevant questions are practical:
- Did the system improve detection or response in a measurable way?
- Were the results independently evaluated?
- Could less intrusive measures have achieved a similar result?
- Did the system create additional staffing, maintenance or data-management costs?
- Were there errors, inappropriate uses or unjustified intrusions into privacy?
A serious European security policy should require evidence of performance and regular reassessment. Public authorities should not have to reject a technology simply because it is intrusive; equally, they should not be permitted to deploy it indefinitely simply because it is available.
- Brussels needs a better public evidence base
Brussels is a particularly relevant setting for this debate. The capital hosts demonstrations, political events and large public gatherings alongside the ordinary activity of an international city. Belgian Federal Police statistics allow crime patterns to be explored at local levels, but public crime figures alone do not provide a complete assessment of surveillance at demonstrations. [2] A more useful evaluation would connect three types of evidence.
First, incident data. Authorities should publish appropriately aggregated information about public-order incidents at major gatherings, distinguishing documented offences from arrests, allegations and convictions. The methodology should make clear which events and offences are included.
Second, surveillance deployment. Relevant public authorities should make information available on camera networks in public spaces, their stated purposes, the responsible operators and the legal arrangements governing access to recordings. Where publication of exact locations would create a legitimate security risk, appropriately aggregated data could still improve public understanding.
Third, operational outcomes. Authorities should report how often footage materially assisted investigations, whether it contributed to identifying suspects, what limitations arose and how systems are evaluated. Such figures should distinguish an investigative lead from a successful identification, an arrest or a prosecution.
These data should not be interpreted as a simple scoreboard. Changes in recorded crime may reflect multiple factors, including reporting behaviour, policing priorities, event size and wider social conditions. Before-and-after comparisons alone are rarely sufficient to establish causation.
Nevertheless, a transparent evidence base would allow citizens, elected representatives and researchers to ask whether surveillance investments deliver proportionate public benefits.
The objective is not to publish sensitive operational details. It is to ensure that decisions affecting public space can be scrutinised without requiring citizens to accept official assurances on trust alone.
- The invisible side of security
At the GATHERINGS conference, Dr Mehdi Ghassemi of the Université Catholique de Lille explored what he called the aesthetic politics of frictionless security. His presentation examined a vision of security in which people continue moving through a space while sensing, screening and classification take place in the background. [4]
This vision promises convenience. People may face fewer visible interruptions, while security processes operate through technical systems and information flows.
But less visible security does not necessarily mean less intervention. Screening may still lead to additional checks, and operational burdens may be shifted to staff or concentrated on people flagged by a system. Ghassemi’s argument invites us to ask where friction remains, who experiences it and whether the system’s decisions can be challenged. [4] This raises a democratic concern, If a security process becomes less visible, the public may find it harder to understand what information is being collected, how classifications are made or who is responsible for a decision.
The issue is not that every unobtrusive measure is inherently wrong. It is that convenience should not be confused with accountability.
A system may keep a queue moving while making its internal decisions less transparent. A person may be subject to screening without knowing which information triggered it. A private contractor may provide a technology while responsibility for its use is distributed among several public and private actors.
In such circumstances, citizens need clear information about the purpose of surveillance, the rules governing data access and the routes available to question an error.
A democratic society should not have to choose between efficient public services and the ability to scrutinise public power.
- Technology must earn public trust
Dr Ciara Bracken-Roche’s presentation, Balancing Interests in Public Surveillance: Democratic Oversight of Surveillance Technologies for Policing and Security, examined the challenges raised by technologies including facial recognition, automated identification and other forms of police surveillance. Her presentation emphasised the importance of evaluating effectiveness, necessity, proportionality, privacy and possible chilling effects. [5]
These concerns are especially relevant when public authorities make purchasing decisions based on claims about accuracy, speed or reduced staffing requirements.
The risks are not merely theoretical. In November 2024, the US Federal Trade Commission announced action against Evolv Technologies over allegations that the company had made misleading claims about its AI-powered security-screening systems. The FTC complaint alleged, among other things, that the company overstated weapon-detection capabilities and made misleading claims about false alarms, screening speed and labour savings. The proceedings concerned specific products and claims; they do not establish that every AI screening system is ineffective. [6]
The case nevertheless illustrates an important procurement principle: claims about safety-critical technology should be supported by evidence, not accepted because a product is marketed as artificial intelligence. Public purchasers should demand independent testing, clear performance criteria, documentation of limitations and contractual accountability. They should also assess whether a system performs differently across relevant operating conditions and whether its outputs require human verification. Where a system can affect a person’s treatment by authorities, an inaccurate classification is not simply a technical inconvenience. It can lead to unnecessary scrutiny, exclusion or other consequences that deserve procedural safeguards.
Innovation and oversight are therefore not opposing objectives. Reliable evaluation helps distinguish useful technology from expensive promises, protects public resources and supports trust in genuinely effective tools.
- Privacy and the freedom to assemble are not optional
The legal framework is clear that public safety measures must be considered alongside fundamental rights. Articles 7 and 8 of the Charter of Fundamental Rights of the European Union protect respect for private and family life and the right to the protection of personal data. Article 12 protects freedom of peaceful assembly and association. [7]
These rights do not mean that every camera in a public space is unlawful. Nor do they prevent authorities from taking proportionate measures to respond to serious threats. They do require the relevant legal basis, purpose, safeguards and effects of a measure to be considered. Belgium’s Data Protection Authority explains that the Camera Act regulates the use of surveillance cameras. Generally, recorded images may be retained for no longer than one month, subject to legal exceptions, including certain cases where images may help establish an offence, damage or a disturbance of public order, or identify a person involved. Specific categories of sensitive locations may be subject to different maximum periods. [8]
Retention is only one part of the issue. The purpose of collection, access to footage, data transfers, security arrangements and possible reuse also matter. Surveillance can have consequences beyond the immediate collection of images. If people believe that attending a peaceful demonstration will expose them to unnecessary identification or opaque tracking, they may have reasons to reconsider participation. The extent of any chilling effect should be investigated rather than presumed, but the possibility is relevant to a rights-based assessment.
Demonstrations are not simply crowds to be managed. They are one-way citizens make demands, contest decisions and participate in political life.
A democratic Europe should protect the safety of demonstrators, bystanders, journalists, workers and police officers while also preserving the conditions for peaceful collective expression.
- The EU has a role—but rules must be enforceable
The European Union already provides a framework for fundamental rights and data protection, and the AI Act establishes restrictions and safeguards for specific uses of AI. For example, Article 5 of the AI Act generally prohibits real-time remote biometric identification by law enforcement in publicly accessible spaces, subject to narrowly defined exceptions and conditions. Those exceptions include specific purposes set out in the legislation and safeguards involving necessity, proportionality and authorisation under the applicable rules. The exact legal position depends on the system and circumstances; ordinary CCTV and real-time biometric identification are not interchangeable categories. [9]
The existence of legislation is important, but legislation alone does not guarantee good practice. Authorities need the expertise to assess technologies, procurement processes that demand evidence, independent oversight, effective remedies and meaningful transparency.
The GATHERINGS project identified a related challenge: surveillance at public events can involve multiple technologies, stakeholders and data transfers, while professionals and citizens may lack clear information about applicable rights and practices. The project sought to address these issues through common standards, an assessment matrix and awareness-raising activities. [1]
A European approach can help make standards more consistent, encourage the exchange of good practice and support public authorities that lack specialist resources. It should not mean centralizing every operational decision in Brussels. Decisions should be taken at the appropriate level, with European action where common standards, cross-border coordination or the protection of shared rights offer clear added value.
This is a federalist argument for coordinated and accountable governance: shared European principles, effective institutions and democratic scrutiny at every level.
- Four questions before any surveillance system is approved
The GATHERINGS framework provides a useful starting point for public authorities, event organizers and policymakers. Before introducing or expanding a surveillance measure, they should be able to answer four questions.
Security: What specific risk is being addressed, and what evidence demonstrates that the proposed system will improve safety?
Privacy and rights: What information will be collected, who will have access, how long will it be retained, and how can people challenge misuse or error?
Economic cost: What is the purchase, staffing, maintenance and oversight costs, and have less intrusive alternatives been assessed?
Social impact: Could the measure affect trust, participation in public life or communities disproportionately? What safeguards and independent evaluation are needed?
These questions reflect the project’s central effort to balance security, privacy, economic cost and social impact. [1]
The assessment should take place before procurement, not after deployment has become difficult to reverse. It should also be revisited when a system’s purpose, capabilities or data-sharing arrangements change.
A camera network designed to monitor traffic should not automatically become a tool for identifying participants at demonstrations. A system introduced for one event should not be repurposed without assessing the applicable legal basis and safeguards.
The principle is simple: define the purpose, establish necessity, test effectiveness, minimize data, control access and make accountability visible.
- Democratic security is a European project
JEF Belgium describes its mission in terms of a democratic European Federation that protects peace, the rule of law and human rights. Its Political Platform emphasizes democracy, participation, transparency, subsidiarity and fundamental rights. [10]
Applied to surveillance, these principles point towards a clear political position. Europe should have the capacity to protect people from violence and disorder. Public authorities should be able to use appropriate, lawful and effective tools to respond to genuine threats. But those tools must remain subject to democratic control, and their benefits must be weighed against their financial, social and rights-related costs.
The alternative is not simply a less secure society. It is the risk of building systems whose purposes expand, whose performance is insufficiently tested and whose decisions become harder for citizens to understand or challenge. A stronger Europe should not be measured by how much it can observe. It should be measured by how effectively it protects people while preserving their dignity, freedom and ability to participate in public life. The right question is therefore not whether Europe should choose security or freedom. It is whether Europe can build security that deserves public trust.
The goal should be neither surveillance for its own sake nor the rejection of every security technology. It should be effective, proportionate and accountable security under democratic control.
That is a debate young Europeans should help lead.
Sources
[1] European Commission, CORDIS, GATHERINGS: Common Standards for Security, Privacy and Cost of the Surveillance of Public Gatherings. Project description, objectives and reporting: https://cordis.europa.eu/project/id/101121200 and https://cordis.europa.eu/project/id/101121200/reporting
[2] Belgian Federal Police, Criminality statistics. Official data, reports and geographical breakdowns: https://www.police.be/statistiques/fr/criminalite
[3] Campbell Collaboration, Effects of Closed Circuit Television Surveillance on Crime. Systematic review: https://www.campbellcollaboration.org/review/effects-of-closed-circuit-television-surveillance-on-crime/
[4] Dr Mehdi Ghassemi, The Aesthetic Politics of Frictionless Security: Contestability and Surveillance of Public Gatherings. Presentation at the GATHERINGS Final Conference, 16 September 2026.
[5] Dr Ciara Bracken-Roche, Balancing Interests in Public Surveillance: Democratic Oversight of Surveillance Technologies for Policing and Security. Presentation at the GATHERINGS Final Conference, 16 September 2026.
[6] US Federal Trade Commission, FTC Takes Action Against Evolv Technologies for Deceiving Users About Its AI-Powered Security Screening Systems, 12 November 2024: https://www.ftc.gov/news-events/news/press-releases/2024/11/ftc-takes-action-against-evolv-technologies-deceiving-users-about-its-ai-powered-security-screening
[7] Charter of Fundamental Rights of the European Union, Articles 7, 8 and 12: https://eur-lex.europa.eu/eli/treaty/char_2016/oj
[8] Belgian Data Protection Authority, Surveillance Cameras: https://www.dataprotectionauthority.be/citoyen/themes/les-cameras-et-votre-vie-privee/les-cameras-de-surveillance
[9] Regulation (EU) 2024/1689, Artificial Intelligence Act, particularly Article 5: https://eur-lex.europa.eu/eli/reg/2024/1689/oj
[10] Young European Federalists (JEF Europe), About JEF and Political Platform: https://jef.eu/about-jef/ and https://jef.eu/resolution/political-platform/
Editorial note: This article combines official project documentation, legal sources and published research with political analysis and policy recommendations. It does not claim that the conference was organised by JEF Brussels or that JEF has formally adopted the article’s conclusions. The presentations support the descriptions of the speakers’ arguments; the policy proposals are editorial recommendations, not verbatim project findings. The views of this article only reflect the author’s.
