Good practices on AI from NHRIs
National Human Rights Institutions (NHRIs) are essential to protecting and promoting human rights in the digital age. As a bridge between the government and civil society and through their capacity to call upon the cooperation and expertise of other bodies, institutions and organisations, NHRIs are well placed to address the variety of human rights impacts that the development and deployment of AI can cause.
NHRIs’ broad mandate offers various opportunities to work on the human rights impacts of AI. This ranges from monitoring and reporting on impacts to giving policy and legislative advice to governments.
Below you can see examples of good practices from NHRIs working on AI, including some case study videos.
Click the buttons below to browse good practices by NHRI mandate:
Advising on policy and legislative measures
Proactive engagement with the UK’s white paper
From: Equality and Human Rights Commission (Great Britain)
The Equality and Human Rights Commission (EHRC) of Great Britain had proactive, early engagement with the UK Government on the AI White Paper (“A pro-innovation approach to AI regulation White Paper”). In June 2023, they released their consultation response calling for more attention to the impact of AI on human rights and equality. The NHRI’s proactive engagement with the UK’s AI White Paper contributed to improving the human rights and equality references in the paper.
Click to open video player to watch good practice video
EU-Council of Europe Technical Support project to strengthen administrative capacity of NHRIs/equality bodies of Belgium, Finland, and Portugal
From: Unia (Belgium NHRI)
Through this project, Unia is preparing for its role under Article 77 of the EU AI Act. The AI Regulation defines cooperation modalities between existing fundamental rights authorities such as Unia and relevant market regulators to ensure that the fundamental rights of individuals are effectively protected against the potential risks of AI. The project aims to increase Unia’s readiness to prevent and address discrimination in public administration’s deployment of AI and automated decision-making (ADM), including by providing legal assistance to those who are discriminated against by AI, issuing policy recommendations, and providing training in line with the EU and Council of Europe standards and national legislation.
Contributing to an expert group on tech giants
From: Danish Institute for Human Rights
The Danish Government has set up an external expert group on tech giants to support the Government’s response to the democratic and fundamental rights issues associated with the tech giants’ practices in a national and European perspective. The expert group presents reports and recommendations to the Government, and has issued, for example, a report on the challenges associated with the tech giants’ development and use of artificial intelligence and how to counter these issues. The report addresses, among other things, the need for issuing additional guidelines and developing good practices for responsible procurement and use of artificial intelligence within the public sector. It also looks at the duty of tech giants to address the systematic risks posed by the proliferation of harmful and misleading content generated by artificial intelligence on their platforms. Senior Researcher Rikke Frank Jørgensen, from the Danish Institute for Human Rights (DIHR), serves on the group as appointed expert. The reports of the expert group can be found here.
Engagement with the French Parliament to prevent the application of automated video surveillance
From: French National Consultative Commission on Human Rights
The French National Consultative Commission on Human Rights (CNCDH) works on the topic of AI and securitisation. In December 2022, the French government tabled a draft law relating to the organisation of the Olympic Games in the summer of 2024 in France. This text included an experimentation of the automated video surveillance (AVS) to contribute to the event’s security. CNCDH engaged with the French Parliament to prevent this application of automated video surveillance by sending a letter to the deputies to alert them to numerous concerning provisions (lack of information and previous evaluation of this new technology and the wide scope of the AVS experiment). The deputies, however, adopted the text, adding few guarantees.
Collaboration with the Danish Institute for Human Rights and Council of Europe to build capacity in digitalisation and AI
From: People’s Advocate of Moldova
The People’s Advocate of Moldova is strengthening its digitalisation and AI capacities through a collaboration with the Danish Institute for Human Rights. Key actions include assessing the impact of digitalisation on human rights and enhancing the People’s Advocate’s capacity to prevent and address human rights violations in the digital sphere, including through policy analysis, monitoring mechanisms, and evaluating measures adopted in the field. In partnership with the Council of Europe, the People’s Advocate is evaluating AI tools for integration into its operations, including a potential chatbot that could help direct individuals to the appropriate specialist based on their issues and provide useful information about the NHRI’s activities.
Shaping the framework for the use of AI in the public sector
From: Netherlands Institute for Human Rights
The Dutch government is developing a framework for the use of AI in the public sector. In the early stages, the Netherlands Institute for Human Rights engaged with government officials and provided feedback on the first draft. The framework has since evolved with input from various stakeholders, who participated in both drafting and advisory groups. The Institute is part of the advisory group on “bias and non-discrimination”.
Contributing to government position on Council of Europe recommendation on AI, equality, and discrimination
From: Norwegian Human Rights Institution
The Norwegian Human Rights Institution is part of a working group established by the Norwegian government to help shape its position as a member of the Council of Europe’s Committee of Experts on Artificial Intelligence, Equality and Discrimination, that is tasked to draft a Recommendation on the impact of artificial intelligence systems, their potential for promoting equality – including gender equality – and the risks they may cause in relation to non-discrimination. The working group is composed of experts with practical and legal experience in the field of AI, with the NHRI contributing human rights expertise. The Institution provides feedback on the draft and both general and technical input to the government’s potential position. To date, many of the group’s suggestions have been incorporated into the government’s final position.
Addressing Issues with the BOSCO Algorithm
From: Ombudsman of Spain
The Ombudsman of Spain has actively addressed issues with the BOSCO algorithm, a system provided to utility companies by the Ministry for Ecological Transition and Demographic Challenge, to identify who qualifies as a vulnerable consumer for subsidised electricity supply. This algorithm, which relies on income data from the Spanish Tax Authorities, failed to recognise applicants below a certain income threshold who did not file income tax statements. As a result, many vulnerable consumers were wrongly excluded from assistance without a clear explanation or the opportunity to rectify their application. There is no direct link between vulnerable consumers and the Administration, worsening their situation. No fast procedure exists to verify compliance within the algorithm-based system. The remedy for rejection – issued by a company – is handled by consumer authorities, despite relying on Administration-held data. Many beneficiaries are unaware of this rigid process, delaying potential exclusion. Additionally, electronic proceedings may be limited or inaccessible to the most vulnerable applicants. In response, the Ombudsman recommended that the Ministry for Ecological Transition and Demographic Challenge amends regulations to allow applicants to submit additional documentation on their income level directly to the relevant energy authorities, speeding up the verification process.
Recommendation for the adoption of the Council of Europe AI Convention
From: Human Rights and Equality Institution of Türkiye
The Human Rights and Equality Institution of Türkiye submitted an opinion letter to relevant public institutions regarding the Council of Europe Framework Convention on Artificial Intelligence, Human Rights, Democracy and the Rule of Law, stating that Türkiye’s signature would be a positive step toward the country’s commitment to protect human rights, democracy, and the rule of law. It noted that this would confirm the determination to increase the reliability of artificial intelligence systems at the national level. It would also provide positive outputs by strengthening the strategic goal of establishing an appropriate ethical and legal framework for the field of artificial intelligence.
Developments on AI regulation
From: Ukrainian Parliament Commissioner for Human Rights
As part of the EU4DigitalUA project, the Ukrainian Parliament Commissioner for Human Rights collaborated with the Ministry of Digital Transformation of Ukraine to develop the methodological material “Human Rights in the Age of Artificial Intelligence: Challenges and Legal Regulation”. It also participated in the Expert and Advisory Committee on the Development of AI in Ukraine under the Ministry of Digital Transformation and plans to join a working group for research on high-tech solutions using the Sandbox method. Additionally, the Secretariat of the Ukrainian Parliament Commissioner for Human Rights is involved in the working group of the National Association of Advocates of Ukraine on the legal regulation of artificial intelligence. While the NHRI is tasked with overseeing human rights compliance, it also publishes informational materials on official web resources and conducts educational events on human rights protection in the context of AI technologies. These initiatives are aimed at government authorities, National Police of Ukraine personnel, and civil society representatives.
Awareness raising, education and outreach
Learning resource on digital rights for NHRIs
From: Danish Institute for Human Rights
The Danish Institute for Human Rights has published a Digital Rights resource for NHRIs that was co-created with members of the NHRI Digital Rights Alliance. The learning resource explores the impacts of technology on human rights and the role of NHRIs in promoting and protecting human rights both online, offline and somewhere in between.
MOOC on Human Rights Based Approach to AI
From: Danish Institute for Human Rights
The Danish Institute for Human Rights, together with the Global Campus on Human Rights, Queens University Belfast and the Raoul Wallenberg Institute, has co-developed a new Massive Open Online Course (MOOC) on a Human Rights Based Approach to AI. The MOOC explores the profound impact AI can have on human rights, both positively and negatively, by presenting new challenges for individuals, communities, and society as a whole. Through diverse case studies, the Institute illustrates how adopting a Human Rights Based Approach (HRBA) can help navigate both the promises and perils of AI, highlighting why this framework is crucial in addressing these complex issues. The course is open to all, free of charge and targeted at a general audience.
Examination of the impact of the use of biometrics and data exchange practices
From: German Institute for Human Rights
The German Institute for Human Rights has raised awareness on digital rights, especially in the context of the use of AI within the asylum and migration sector. For many years, the fields of refugee protection and data protection had few points of contact in Germany. Against the backdrop of the growing collection of refugees’ data and the proliferation of digital tools to process asylum claims in the wake of 2015, the German Institute for Human Rights, in collaboration with Caritas Germany, convened an expert discussion in June 2019 entitled “Transparent Refugees? The Human Right to Privacy in Asylum Procedures and Integration Measures.” The event brought together around 30 experts from human rights and refugee organisations, charities, privacy advocacy groups, data protection authorities, federal ministries, and immigration authorities to discuss issues of privacy and data protection in the context of asylum procedures and migration management.
Conference on digitisation and AI
From: People’s Advocate of Moldova
In September 2024, the People’s Advocate of Moldova organised an international conference on “Human rights protection mechanisms and the role of National Human Rights Institutions in the European Union and beyond”, which included a special panel on the role of NHRIs in digitalisation, AI and human rights. It sparked rich discussions on topics such as risks to human rights in digital environments and the use of AI for human rights protection. In addition, leading organisations such as the Council of Europe, UNESCO and the EU institutions shared best practices and frameworks.
Strategic development program for integrating human rights in the digital sphere
From: People’s Advocate of Moldova
The People’s Advocate of Moldova is implementing its Strategic Development Program (2023–2030), focusing on enhancing the integration of human rights in the digital sphere to prevent exclusion. Key actions include analyzing public documents and advocating for digitalization legislation from a human rights perspective, monitoring public policy commitments to effectively protect children in online environments, developing and disseminating guidelines to ensure human rights in the implementation of digital systems, and identifying and conducting in-depth examinations of strategic cases involving human rights violations.
Organisation of a capacity building workshop on the Human Rights Impact Assessment (HRIA) in the development of technologies and AI systems for human rights experts
From: Human Rights and Equality Institution of Türkiye
The Human Rights and Equality Institution of Türkiye (HREIT) organised a capacity building workshop on the Human Rights Impact Assessment (HRIA) in the Development of Technologies and AI Systems for human rights experts and Assistant Experts on 1-2 February 2024. The training was divided into five building blocks (human rights and AI; the risk-based approach; HRIA; the HRIA model and function; and case studies) and aimed to raise participants’ awareness of the challenges and opportunities in promoting human rights-based approaches in AI through the development and use of HRIAs.
Call for book chapters for an upcoming academic book project titled “Digitalisation and women’s rights”
From: Human Rights and Equality Institution of Türkiye
The Human Rights and Equality Institution of Türkiye (HREIT) published a call for book chapters on their website and social media accounts on 29 February 2024 for an upcoming academic book project (in English) titled “Digitalization and Women’s Rights”. One of the book’s subchapters is devoted to “Women and Artificial Intelligence Implementations”. As a result of the intense interest and many submissions, the scope of the book has been expanded and it will be published in two volumes instead of one. The book is expected to be published by the end of March 2025.
Handling complaints, conducting investigations and strategic litigation
Biometric data case
From: Ombudswoman of the Republic of Croatia
Based on information from the media, the Ombudswoman of the Republic of Croatia has found that Bolt implemented a system that collects and processes biometric data of the driver-partner (selfies) for the purpose of identity verification. The Bolt app asks drivers to take selfies, after which the photo is compared to the one on their driver’s license. The whole procedure has to be repeated regularly at the start of each driving service. If drivers do not provide selfies or refuse to verify them, they are not allowed to continue the service. The Ombudswoman opened an inquiry procedure and asked Bolt for additional information. The Data Protection Agency is also conducting a supervisory procedure regarding biometrics. The investigation is ongoing.
Click to open video player to watch good practice video
The right to privacy of platform workers
From: Ombudswoman of the Republic of Croatia
The Ombudswoman of the Republic of Croatia initiated an inquiry procedure regarding the possible violation of the right to privacy of platform workers (Bolt, Glovo, Wolt, Uber). The Workers’ Union informed them that apps collect location data even when the workers are not active on the platform. They asked a few digital platforms to inform them on different aspects of how the app works (private phone or official mobile phone, what data is collected, and for what purpose, requirements regarding workers’ presence on the app, sanctions in case of the exclusion of workers from the app and possible remedies). The investigation is ongoing.
Click to open video player to watch good practice video
Right to privacy regarding the access to citizens’ health data
From: Ombudswoman of the Republic of Croatia
The Ombudswoman of the Republic of Croatia initiated a procedure regarding the potential violation of the right to privacy regarding the access to citizens’ health data contained in the electronic e-health file system. The Ombudswoman requested that the Data Protection Agency monitor the categories of healthcare professionals who have access to citizens’ health data contained in the e-health file. The investigation is ongoing.
Click to open video player to watch good practice video
Launch of the legal clinic “AI and human rights”
From: Unia (Belgium NHRI)
AI is increasingly used in situations that can have a significant impact on people’s lives. It may lead to violations of human rights, such as the right to non-discrimination, the right to privacy or the right to an effective remedy. Unia has encountered difficulties in identifying and addressing cases of algorithmic discrimination. Due to the power and information asymmetry in AI-related cases, building strong legal arguments is time-consuming and requires substantial expertise. In 2023, Unia launched a legal clinic in collaboration with the renowned Faculty of Law at KU Leuven, under the guidance of Professor Nathalie Smuha. The team of coaches consists of people with both legal and technical expertise and guides students who have proven invaluable in supporting Unia’s legal proceedings.
Manjang vs Uber Eats Case
From: Equality and Human Rights Commission (Great Britain)
The Equality and Human Rights Commission (EHRC) of Great Britain together with the App-Drivers and Couriers Union (ADCU) provided funding for the case of Manjang. Mr Manjang, an Uber Eats driver, filed a complaint in 2021 against the Uber Eats app for being racially discriminatory. During the case, the EHRC and the ADCU raised concerns about the Uber Eats app’s use of AI and automated processes and its ability to suspend drivers’ access to the app, depriving them of their income, and its lack of transparency. The case was settled in March 2024 with a financial settlement for Mr Manjang, plus restoring of his access to the app.
Assessment of Facebook’s advertising algorithm
From: Netherlands Institute for Human Rights
Stichting Clara Wichmann and Global Witness requested the Netherlands Institute for Human Rights to assess whether Facebook’s advertising algorithm discriminates by showing job ads for stereotypically male and female positions mainly to the relevant group. Meta confirmed its algorithm uses gender as a criterion and learns from user clicks. The Institute found that Meta, as a social media platform, is responsible for monitoring its algorithm’s functioning and investigating potential stereotyping. It should evaluate the data used for training and the impact of algorithmic selections on different groups. Since Meta Ireland has not demonstrated sufficient action to neutralise stereotyping reinforcement, the Netherlands Institute for Human Rights concluded that the algorithm discriminates on the grounds of gender.
Assessment of discrimination in banking
From: Netherlands Institute for Human Rights
In July 2024, the Netherlands Institute for Human Rights addressed complaints from two clients of ING, a Dutch bank, who claimed their payments were blocked and verified due to having “non-Dutch surnames”. ING’s automated system, used to detect suspicious transactions under the Dutch Anti-Money Laundering and Anti-Terrorist Financing Act, screens transactions for matches with names on both national and international sanction lists, many of which include individuals with non-Dutch names. Although ING does not explicitly target non-Dutch names, the system disproportionately impacts persons with non-Dutch names. The Institute found that this method resulted in indirect discrimination based on non-Dutch origin (see decisions in Dutch here and here). Following media attention, the NHRI reviewed similar complaints. In December 2024, it found ING had indirectly discriminated on the grounds of race/nationality against a general practitioner receiving a payment. ING’s system flagged the payment due to a possible link to Iran, and despite the payer being a passer-by, ING requested further documentation. The Institute ruled this second request unjustified and thus discriminatory. In contrast, no discrimination was found in a Rabobank case, where extra information was requested from a client with Dutch and Russian nationality after a payment to her Turkish partner. The Institute found Rabobank’s actions were based on geographic risk, not race/nationality.
Dealing with AI-related cases
From: Netherlands Institute for Human Rights
Several AI-related cases have been brought before the Netherlands Institute for Human Rights, which also serves as a National Equality Body. These include 1) multiple cases concerning childcare allowance, where the Institute concluded that there was a presumption of discrimination against individuals with a foreign origin, leading to a reversal of the burden of proof and eventually to the conclusion that some individuals were indirectly discriminated against; 2) a case concerning a dating app that needed to be modified to ensure equality as the algorithm was discriminatory; and 3) a case concerning anti-cheating software thought to be discriminating against a student with darker skin.
Monitoring and reporting on AI and human rights
Research into the Albanian digital transformation
From: People’s Advocate of Albania
In 2022, Albania’s government suddenly decided to transition 95% of its public administration services to online-only delivery (e-Albania portal). This sudden change is problematic, as many Albanians lack access to this portal due to limited economic resources or low digital skills, high risk of data leakage and blocked online services and vulnerable groups facing barriers to accessing online services. The People’s Advocate, with the support of ENNHRI, therefore initiated the implementation of a project aimed at assessing the impact of the digitisation of public services. Through this project, the People’s Advocate aimed to conduct an in-depth analysis of the impact of digitalising public services in Albania, culminating in the drafting of a special report that was published in September 2024.
Click to open video player to watch good practice video
Development of the ‘Guidance on Human Rights Impact Assessments on Digital Activities’
From: Danish Institute for Human Rights
The Danish Institute for Human Rights (DIHR) developed the Guidance on Human Rights Impact Assessments on Digital Activities document which provides an accessible and practical means to guide human rights practitioners, businesses and public authorities, as well as other actors interfacing with the digital ecosystem, to conduct human rights impact assessments (HRIAs). The tool is targeted towards identifying, examining and addressing human rights impacts on digital activities. The HRIA methodology and process is also summarized in Key Principles for HRIA of digital business activities. Another tool developed by the DIHR is the Digital Rights Check, a web-based tool to assist those working on development projects to assess the potential human rights impacts of their digital projects or projects components. A third paper developed by DIHR is on public procurement as a tool to address human rights risks in the use of digital technology to deliver essential public services. It considers how policy makers and public procurement professionals can ensure that the public procurement process identifies human rights risks linked to the application of digital technologies and has measures in place to mitigate such risks before harm occurs.
Click to open video player to watch good practice video
Contribution to the Impact Assessment Mensenrechten en Algoritmes (IAMA) and the Handbook ‘Non-discrimination by design’
From: Netherlands Institute for Human Rights
The Netherlands Institute for Human Rights has one of its strategic programme with a focus area within digitalisation and human rights. They contributed to the Handbook ‘Non-discrimination by design’, which was commissioned by the Dutch government and aims to help both the public and private sector prevent discriminating effects of the AI system they (want to) use. It provides an overview of the technical, legal and organisational perspectives that should be taken into account when developing and implementing an AI system. Another helpful tool is the Impact Assessment Mensenrechten en Algoritmes (IAMA), an overall AI lifecycle impact assessment model relating to the development and deployment of AI systems which includes a focus on human rights. This tool is developed specifically for the public sector. It includes a section to identify and address biases and assumptions in the dataset and calls for parties to ensure complete and accurate datasets. The purpose of the IAMA is to assist in deciding whether or not an algorithmic application should be developed, and if so, to then responsibly implement the chosen development. The IAMA was commissioned by the Dutch government and created by Utrecht University. The Institute endorses the IAMA and promotes its use through publications on its website and communications with the government.
Monitoring of the use of new technologies within operational policing
From: Scottish Human Rights Commission
In 2021, the Scottish Human Rights Commission issued a Paper on ‘Human Rights and New Technologies in Policing’, an independent analysis of the human rights issues linked to new and emerging technologies in policing. It was produced as a contribution to support the work of the Independent Advisory Group on Emerging Technologies in Policing, of which the Commission was a member.
Visit the other pages on the AI resource
