Individual freedoms refer to the set of rights that protect each person against abusive interference from the state or third parties: freedom of expression, right to privacy, freedom of movement, protection against arbitrary surveillance. These rights, enshrined in the Declaration of the Rights of Man and of the Citizen and the European Convention on Human Rights, form the legal foundation upon which life in democracy rests.
Their defense takes on a new dimension as control technologies advance and security legislations multiply. Understanding the concrete mechanisms that threaten these freedoms allows for better protection.
Artificial Intelligence and Freedoms: What the AI Act Prohibits Since 2025
The entry into force of the European AI Act has shifted the defense of individual freedoms onto technical ground. Published in the Official Journal of the EU on July 12, 2024, and effective from August 1, 2024, this regulation classifies certain uses of artificial intelligence as presenting an “unacceptable” risk to fundamental rights.
Since February 2, 2025, several AI practices are formally prohibited in Europe:
- Social scoring, which would involve evaluating and ranking citizens based on their behavior, similar to the social credit system experimented with in other countries.
- Remote biometric identification in real-time in public spaces for law enforcement purposes, except under strictly regulated exceptions.
- Mass data extraction (web, video surveillance) to create facial recognition databases, as well as biometric categorization aimed at deducing protected characteristics such as ethnic origin or political orientation.
- Emotion recognition in the workplace and educational institutions, and exploiting an individual’s vulnerabilities (age, disability) by an automated system.
This regulatory framework marks a turning point. The protection of freedoms no longer relies solely on constitutional texts or the jurisprudence of the Council of State: it now plays out in the technical architecture of the systems deployed daily. Resources document this evolution, notably on https://www.noslibertes.org/, which gathers analyses and monitoring tools on the subject.
Digital Surveillance and the Right to Privacy: Increasingly Blurred Boundaries
Facial recognition illustrates the ongoing tension between public security and privacy protection. In London, facial recognition cameras have been deployed in the subway. In France, this type of device faces a stricter legal framework, but the pressure to adopt it exists.
The right to privacy is a constitutional right, protected both by Article 8 of the European Convention on Human Rights and by the jurisprudence of the French Constitutional Council. Any infringement of this right must be justified and proportionate, a principle that the administrative judge controls on a case-by-case basis.
The Council of State regularly reminds of this requirement. In its decisions, it checks that national security or law enforcement measures do not disproportionately infringe on freedoms. The proportionality control constitutes the central legal mechanism: the end does not justify all means, even when the stated objective is the protection of public order.
Algorithmic Video Surveillance: A Concrete Case
Algorithmic video surveillance, which automatically analyzes video streams to detect “suspicious behaviors,” poses a fundamental problem. The system does not merely record: it interprets. This layer of automated analysis transforms a passive camera into a behavioral prediction tool.
The risk, documented by the CNIL, lies in the opacity of these systems. A citizen filmed in a public space does not know if their behavior is being analyzed, classified, or based on what criteria. This asymmetry of information between the state and the individual undermines the very principle of informed consent.
Freedom of Expression and Regulation of Digital Platforms
The Digital Services Act (DSA), which has come into effect in the European Union, imposes transparency obligations on major digital platforms regarding their moderation and recommendation algorithms. The stated goal is to combat misinformation and illegal content.
The difficulty lies in the demarcation line between legitimate moderation and censorship. Removing illegal content falls under the law, but filtering opinions falls under freedom of expression. Automatic moderation algorithms do not always make this distinction. Journalistic, satirical, or activist content is sometimes removed without human review, with limited appeal procedures.
In France, the Constitutional Council recently examined several legislative provisions affecting digital freedoms. The Ripost law, largely validated by the Constitutional Council in August 2026, contains measures related to the regulation of certain online practices. Each new security law undergoes a constitutional review that assesses whether the restrictions on freedoms are necessary, appropriate, and proportionate.
Algorithms and Information Lock-in
Beyond content removal, recommendation algorithms shape access to information. By directing each user towards content that aligns with their detected preferences, they reduce the diversity of viewpoints to which a person is exposed.
This mechanism, often referred to as a “filter bubble,” is not neutral for the exercise of citizenship. An effective freedom of expression requires pluralistic access to information. The regulation of recommendation algorithms, as provided by the DSA, attempts to address this challenge, but its concrete implementation remains to be evaluated.
Protection of Personal Data: The Role of the GDPR and Its Limits
The General Data Protection Regulation (GDPR) has been the European reference framework for personal data protection since 2018. It enshrines several concrete rights for individuals: right of access, right to rectification, right to erasure, right to data portability.
Its application reveals structural limits. The sanctions imposed on companies that violate the regulation are sometimes insufficient compared to the economic benefits derived from data exploitation. The imbalance between individuals’ control capacity and the data collection power of digital actors remains an unresolved issue.
The interplay between the GDPR and the AI Act creates a double layer of protection: the former regulates the collection and processing of data, while the latter regulates the automated uses that arise from it. This complementarity is recent, and its effectiveness will largely depend on the ability of supervisory authorities, such as the CNIL in France, to exercise effective oversight.
The defense of individual freedoms is not limited to an abstract debate on democratic values. It plays out in precise technical provisions: parameters of a facial recognition algorithm, moderation criteria of a platform, conditions for retaining biometric data. Every choice of digital architecture is a societal choice, and it is at this level of detail that collective vigilance remains crucial.