Back to General issues

 Quick links

• Effective protection of rights 
• Subsidiarity and the margin of appreciation
• Evolutive Interpretation and the ‘Living Instrument’ Doctrine
• Positive Obligations
• Human Dignity
• Personal Autonomy and Self-Determination
• Lawfulness, Legitimate Aim, Necessity, Proportionality, and Fair Balance

Effective Protection of Rights

The ECHR and the ESC are intended to guarantee rights that are not merely theoretical or illusory but practical and effective.[1] National authorities must ensure that rights holders can effectively enjoy their rights, which may involve adopting legislation, ensuring its effective application, providing adequate resources, and establishing appropriate operational procedures. Accordingly, States should safeguard the effective protection of human rights against harms related to activities within the lifecycle of AI systems through measures which may include adopting or implementing laws, providing resources, establishing, or designating existing national human rights structures, such as national human rights institutions (NHRIs), as independent oversight mechanisms, and ensuring effective cooperation between such mechanisms and other national human rights structures.
Back to top

Subsidiarity and the margin of appreciation

Subsidiarity means that the States bear the primary responsibility to secure to everyone within their jurisdiction the rights and freedoms defined in the ECHR.[2] The Court authoritatively interprets the ECHR and acts as a safeguard for individuals whose rights and freedoms are not secured at the national level.[3]

National authorities may enjoy a “margin of appreciation” in how they apply and implement the ECHR, depending on the circumstances of the case and the rights and freedoms engaged. This reflects that the ECHR system is subsidiary to the safeguarding of human rights at national level and that national authorities are in principle better placed than an international court to evaluate local needs and conditions.[4] Under the ESC, States Parties also have discretion in determining the steps to comply with its provisions, balancing general interests with the needs of specific groups and available resources. With respect to new technologies, in particular, any State claiming a pioneer role in their development bears special responsibility for striking the right balance between the potential benefits of their extensive use against protected rights.^[5]
Back to top

Evolutive Interpretation and the ‘Living Instrument’ Doctrine

The ECHR and the ESC are "living instruments", interpreted dynamically in the light of present-day conditions to address evolving societal and technological issues.[6] The Court’s past rulings on issues like data interception,[7] biometric data,[8] the internet and digital tools,[9] or facial recognition technology^[10] highlight its capacity to adapt the application of existing rights to modern challenges. Likewise, the ECSR has addressed the right to privacy in the context of emerging new technologies.[11] By applying this doctrine, both the Court and the ECSR are expected to apply the ECHR and the ESC to AI-related cases in the future.
Back to top

Positive Obligations

States have a duty under both the ECHR and the ESC to refrain from unjustified interference with human rights (“negative obligations”) and to ensure their effective realisation and protection (“positive obligations”). Substantive positive obligations require the basic measures needed for full enjoyment of the rights guaranteed (e.g., proper rules governing intervention by the police or prohibiting ill-treatment). Procedural positive obligations require domestic procedures to ensure the protection of rights holders (e.g. conducting an effective investigation).  

Positive obligations can apply even in cases where threats originate from private individuals or entities beyond direct state control as these instruments can address both vertical relationships – between national authorities and individuals – and horizontal relationships[12], between individuals or entities. States may be required to protect human rights in the sphere of the relations between individuals themselves (horizontal effect). This duty becomes particularly important in the context of the deployment of AI systems, where public-private partnerships and procurement from private actors are prevalent.

States must act diligently and reasonably, taking appropriate measures within their resources and capacities. Positive obligations may require the State to ensure the existence of adequate and effective mechanisms under which sanctions may be imposed in particular cases, enact specific legal rules, and/or take operational steps to protect individuals from foreseeable risks to their rights.[13]

States’ positive obligations[14] therefore may require them to take the necessary measures to safeguard human rights, including where necessary to assess the relevant risks and potential impacts to human rights that may be posed by AI systems and consider measures to address those potential harms effectively, and/or to implement measures to mitigate identified risks. The Framework Convention contains a dedicated provision prescribing the need to identify, assess, prevent and mitigate ex ante and, as appropriate, iteratively throughout the lifecycle of the AI system the relevant risks and potential impacts to human rights, democracy and the rule of law by following and enabling the development of a methodology with concrete and objective criteria for such assessments.[15]
Back to top

Human Dignity

 In the ECHR system, human dignity is invoked by the Court to affirm individuals' intrinsic worth and equality.[16] The Court has held that “[r]espect for human dignity forms part of the very essence of the Convention”.[17] The ESC system too recognises that human dignity is the fundamental value and indeed the core of positive European human rights law – whether under the European Social Charter or under the European Convention of Human Rights.[18]

The Framework Convention also requires that the respect for human dignity be among the principles that govern artificial intelligence.[19] Activities within the AI lifecycle must not dehumanise individuals, undermine their autonomy, or reduce them to data points, and AI should not be anthropomorphised in ways that infringe on human dignity.[20
Back to top

Personal Autonomy and Self-Determination

Personal autonomy is an important principle underlying the interpretation of ECHR guarantees.[21] It is an important aspect of human dignity and refers to the capacity of individuals for self-determination; that is, their ability to make choices and decisions, including without coercion, and live their lives freely. In the context of AI, individual autonomy requires that individuals have control over the use and impact of AI technologies in their lives, and that their agency and autonomy are not thereby diminished.[22] The Framework Convention also specifically requires that the respect for individual autonomy is among the principles that govern AI.[23]
Back to top

Lawfulness, Legitimate Aim, Necessity, Proportionality, and Fair Balance

Certain ECHR rights are absolute and cannot be subject to derogations in times of emergency, exceptions, or permissible interference. However, States Parties are allowed to restrict certain rights in the ECHR[24] and the ESC,[25] known as “qualified rights”, which allow for a balance between individual and general interests. There are some general requirements in both the ECHR and the ESC which must be satisfied for an interference to be justified. The interference must be (i) ‘prescribed by law’ or ‘in accordance with the law’ (requirement of lawfulness).[26] This means that it must have a clear basis in domestic law, ensuring it is rooted in established legal frameworks. Additionally, the legal basis must be accessible to the public, meaning individuals can know and understand the laws that affect their rights.[27] The interference must also be foreseeable, allowing people to anticipate how and when their rights might be restricted.[28] Lastly, it must be free from arbitrariness and implemented with proper procedural safeguards to ensure fairness and due care.[29] The interference with the right must (ii) pursue a legitimate aim[30] and it must be (iii) necessary (in a democratic society) to achieve the legitimate aim pursued.[31] The test of necessity requires that the limitation is proportionate to the legitimate aim pursued, responds to a pressing social need, and uses the least restrictive means.

States will have to ensure that any restrictions on qualified ECHR or ESC rights resulting from activities within the AI systems lifecycle satisfy all of these requirements.

Back to General issues