27 September 2023
In another warning on the dangers posed by the EU's Artificial Intelligence Act, more than 60 organisations (including Statewatch) are calling on EU legislators to ensure that the text upholds the rule of law. An open letter calls: for fundamental rights impact assessments for all high-risk AI systems; to ensure that systems used for national security purposes fall under the Act; and to ensure that AI developers cannot exempt themselves from the Act. "As the EU navigates the complexities of the digital age, it is of the utmost importance that we do not lose sight of our core values," says the letter.
Support our work: become a Friend of Statewatch from as little as £1/€1 per month.
The open letter was drafted and coordinated by the Civil Liberties Union for Europe (Liberties), the European Civic Forum (ECF), and the European Center for Not-for-Profit Law (ECNL).
The European Union’s bodies are currently engaging in the final stage of negotiations on the Artificial Intelligence Act (AI Act). As representatives of civil society organizations committed to upholding and monitoring the rule of law and democracy in the European Union, we call on legislators to ensure that the AI Act is fully coherent with rule of law standards, including transparency, accountability, and access to justice. Given the current decline of the rule of law in Europe, and the heightened concern over AI’s misuse, the Union cannot afford to miss this crucial opportunity and should introduce strong safeguards to protect the full respect of human rights and the rule of law.
Artificial Intelligence (AI) is rapidly transforming our world. Its influence permeates every facet of our lives, from healthcare and education to governance and law enforcement. While the potential benefits of AI are undeniable, it is imperative that we prioritize the protection of our shared values as laid out in the Treaties of the EU and the EU Charter of Fundamental Rights.
The principle of the rule of law, enshrined in Article 2 TEU, is fundamental to the EU. This core value is a precondition for the realization of fundamental rights and other fundamental values. Rule of law requirements mean, among others: transparent, accountable, democratic, and pluralistic law-making process; legal certainty; prohibition of arbitrariness of the executive powers; effective and equal judicial protection, including access to justice, by independent and impartial courts; separation of powers; and non-discrimination and equality before the law. The rule of law is inextricably linked with the protection of democracy and fundamental rights; the three criteria reinforce each other and therefore the link between the development and deployment of AI systems and the rule of law must be established.
As AI becomes increasingly deployed by both the private and public sectors, the rule of law requires the EU to adopt robust safeguards to protect the very foundation our Union stands on. The misuse of AI systems, including opaque and unaccountable deployment of AI systems by public authorities, poses a serious threat to the rule of law and democracy. In member states where these are teetering, regulatory loopholes, such as national security and law enforcement exemptions, could be exploited to weaken democratic institutions and processes and the rule of law. The AI Act needs to create a robust, secure regulatory environment grounded in the protection of fundamental rights and rule of law.
A central safeguard for the accountable and rights-respecting exercise of power is the mandatory inclusion of fundamental rights impact assessments (FRIAs), as it appears in the European Parliament’s position on the AI Act. We argue that, similarly to the obligation set out in the Digital Services Act, which requires very large online platforms to conduct risk assessments, AI FRIAs should be considered in the context of an emerging trend toward the development of standards and certifications for responsible AI governance practices. FRIAs should be an obligation for all high-risk AI technologies to ensure that they are deployed in a way that upholds the principles of justice, accountability, and fairness. FRIAs provide a structured framework to assess and avoid potential fundamental rights violations, ensuring that AI technologies respect, promote, and protect these rights.
As states are responsible for the proper implementation of the rule of law framework, that public authorities, including law enforcement, conduct FRIAs is not just a recommendation but a necessary safeguard to ensure that AI systems are designed and deployed in full accordance with the values of the EU and the EU Charter of Fundamental Rights.
Transparency, a cornerstone of democratic governance and a key feature to ensure that rules are made and applied in a clear and predictable manner, as the rule of law requires, must be integral to the FRIA process. The findings and outcomes of FRIAs must be publicly accessible, including for the persons affected by decisions made with the use of the AI systems, and open to feedback and review. This ensures that AI deployers remain accountable to the people affected by the use of the systems and provides a platform for civil society organizations, watchdogs, and the general public to engage in constructive oversight.
This accountability framework can only safeguard the rule of law and fundamental rights if it applies to all high-risk AI systems. Any loopholes or exemptions would undermine legal certainty and lead to the arbitrary exercise of power, which is fundamentally incoherent with the principle of rule of law.
Through amendments to Article 6 of the AI Act in the Council and Parliament texts, a loophole has been created that gives companies and public authorities alike the power to unilaterally decide that their AI system should be exempted from the AI Act requirements, even though it is intended to be used in one of the high-risk areas, such as law enforcement, justice, elections, or essential public services. If a provider chooses to exempt themselves, then all consequent obligations for deployers of such systems will no longer apply either. As a result, public authorities who then deploy such high-risk systems would arbitrarily escape all AI Act obligations to safeguard people affected by AI from the abuse of their fundamental rights.
For the protection of the rule of law, there should be no blanket national security exemption in the AI Act. While national security could in some instances be considered legitimate grounds for restrictions to fundamental rights, any exceptions from AI Act requirements should be assessed on a case-by-case basis, in line with the EU Charter of Fundamental Rights and the jurisprudence of the CJEU. The risks to fundamental rights are arguably the most severe when AI systems are deployed for national security purposes, given the lack of a commonly accepted definition of national security and the ensuing risk of abuse. We have witnessed this when an invasive surveillance software – Pegasus – was used in some EU member states to spy on journalists, civil society and opposition politicians. From the perspective of rule of law requirements, it is therefore imperative that basic safeguards envisioned in the AI Act are equally respected when AI technologies are developed or used for national security purposes.
Furthermore, individuals wronged by undue or improper use of these systems must have access to effective legal remedies.
We call on EU legislators to mandate FRIAs to be conducted on all high-risk AI systems, in line with the amendments proposed by the European Parliament in Article 29a, and include rules to ensure FRIAs are conducted in an open and transparent manner and their findings subject to public scrutiny.
We call on EU legislators to reject the European Council’s proposed amendment to Article 2 of the AI Act, which aims to exclude AI systems developed or used for national security purposes from the scope of the Act.
We call on EU legislators to return to the original Commission proposal’s version of Article 6(2) of the AI Act, thereby removing newly added loopholes which would give AI developers, be it from the public or the private sector, the power to unilaterally exempt themselves from the safeguards set out in the AI Act.
As the EU navigates the complexities of the digital age, it is of the utmost importance that we do not lose sight of our core values. We call upon EU legislators to heed our requests as a testament to the EU's commitment to protecting the rule of law from AI deployments. The future of our societies depends on the choices we make today.
Europe and International
Centre for Democracy and Technology
Europe Civil Liberties Union for Europe (Liberties)
European Digital Rights (EDRi)
European Center for Not-for-Profit Law (ECNL)
European Civic Forum (ECF)
European Partnership for Democracy (EPD)
European Sex Workers Rights Alliance (ESWA)
International Commission of Jurists Open Future
World Organisation Against Torture
Ligue des droits humains
Bulgarian Center for Not-for-Profit Law
Bulgarian Helsinki Committee
Centre for Peace Studies
Expanse of Gender and Media Culture ‘Common Zone’
Human Rights House Zagreb Politiscope
Estonian Human Rights Centre
Network of Estonian Non-Governmental Organizations
Ligue des droits de l'Homme (LDH)
Ligue de l'enseignement
Maison de l'Europe de Paris
Bundesnetzwerk Bürgerschaftliches Engagement (BBE)
D64 - Zentrum für Digitalen Fortschritt e. V.
Greek Forum of Refugees Homo Digitalis
European House Budapest
Hungarian Civil Liberties Union
Hungarian Helsinki Committee
Ökotárs - Hungarian Environmental Partnership Foundation
Irish Council for Civil Liberties
Associazione Ricreativa Culturale Italiana (ARCI)
Italian Coalition for Civil Liberties and Rights (CILD)
European Movement Italy
Centre for Public Policy PROVIDUS
Human Rights Monitoring Institute
Bits of Freedom
Health Action International
Elektronisk Forpost Norge (EFN)
Moje Państwo Foundation
National Federation of Polish NGOs
Center for Not-for-Profit Law in Romania (CLNR) Expert Forum
Fundatia pentru Dezvoltarea Societatii Civile
Danes je nov dan, Inštitut za druga vprašanja
Lega Center for the Protection of Human Rights and the Environment (PIC)
Civil Rights Defenders
Secret "trilogue" negotiations on the EU's proposed Artificial Intelligence Act are ongoing, and next week MEPs and EU member state representatives will start discussing bans and prohibitions. The week after, decisions are expected on whether to classify the use of AI for migration and security purposes as "high risk" or not. A statement directed at decision-makers and signed by 115 associations and individuals, including Statewatch, calls for strict limits and controls in the AI Act "to prevent harm, protect people from rights violations and provide legal boundaries for authorities to use AI within the confines of the rule of law."
Almost 120 civil society organisations, including Statewatch, are calling on MEPs to close a massive loophole in the proposed Artificial Intelligence Act introduced under pressure from big tech lobbyists. Without changes, the law will allow developers of AI systems to decide whether or not the systems they produce should be considered "high-risk" or not - an obvious invitation for them to decide that they are not, in order to avoid the extra safeguards that the Act should impose.
Thirty civil society organisations, including Statewatch, have published a joint statement calling for the UK government to ensure that it's approach to artificial intelligence upholds fundamental rights and democratic values.
Spotted an error? If you've spotted a problem with this page, just click once to let us know.
Statewatch does not have a corporate view, nor does it seek to create one, the views expressed are those of the author. Statewatch is not responsible for the content of external websites and inclusion of a link does not constitute an endorsement. Registered UK charity number: 1154784. Registered UK company number: 08480724. Registered company name: The Libertarian Research & Education Trust. Registered office: MayDay Rooms, 88 Fleet Street, London EC4Y 1DH. © Statewatch ISSN 1756-851X. Personal usage as private individuals "fair dealing" is allowed. We also welcome links to material on our site. Usage by those working for organisations is allowed only if the organisation holds an appropriate licence from the relevant reprographic rights organisation (eg: Copyright Licensing Agency in the UK) with such usage being subject to the terms and conditions of that licence and to local copyright law.