In simple terms, a law is a rule that regulates what individuals or entities can or cannot do in society.
Laws give rise to “rights” and “obligations.”
A “right” is a legal principle that establishes entitlements (something that is owed to a person) and freedoms (something that a person is allowed to do freely).
An “obligation” is a legal requirement or duty that an individual or entity has a responsibility to carry out or comply with.
These “rights” and “obligations” can arise in a variety of contexts, from employment to business, and from policing to healthcare, and a variety of areas of law regulate conduct across these different domains.
When bringing a case to court, an individual will usually be claiming that their right has been breached by the other party or that the other party has failed to discharge an obligation in relation to the them.
Types of law
The specific rights or obligations that might arise will vary depending on the area of law that applies in a given context.
At supra-national level, the main area of law is International Law. This is the law that governs the rules of relations between countries or states. International law can create obligations for states in a range of areas, including human rights, environment, conflict and international crimes (it is based on international treaties, read more below).
At domestic or country level, the main areas of law are as follows:
Criminal Law: this is law that establishes when an individual, and in some countries a company, will be considered to have committed a crime. In most countries, it will be the responsibility of the police or prosecutors to investigate and bring criminal law cases to court;
Civil Law: this is law that governs relations between private parties (e.g. individuals and companies). For example, civil law can establish rules for remedying a situation where a person has been wronged by another person (e.g. personal injury, injury to property, harassment, defamation). It also regulates the rights and obligations of parties to a legally binding agreement (i.e. contract law). It can also set out rules for solving disputes around employment, family and property;
Public Law: this is an area of law that governs the relationship between the state (i.e. government) and individuals. These laws aim at ensuring that public authorities and government bodies act rationally, reasonably, fairly and lawfully when they exercise their authority. Furthermore, it ensures that public bodies do not act outside their official powers.
What is human rights law?
Human rights law is an area of law that establishes a set of basic principles and norms that are aimed at ensuring the dignity, freedom, and equality of all human beings.
Many of these principles and norms are set out under International Law, and they have the effect of establishing legally binding obligations on countries to ensure that human rights, a number of which are considered on our Spotting There Is A Problem page, are fully respected, promoted and protected within their societies.
The obligations that are owed by countries can be further broken down into:
Negative Obligations: obligations to refrain from doing something that would violate human rights. For example, refrain from censoring people; and
Positive Obligations: obligations to do something to ensure respect and protection of human rights. For example, an obligation to stop someone from violating the human rights of another person.
As international human rights law only formally applies between states, there are limited opportunities for individuals to formally enforce these laws before their country’s courts.
However, since international human rights law gives rise to positive obligations, countries that are bound by this law must ensure the laws and regulations in their own country (see constitution, legislation and case law below), and how they are applied by the courts, sufficiently protect and respect human rights.
This means that countries, in trying to discharge their positive obligation, might put in place human rights-type laws that individuals can then use to take cases against other individuals or entities.
At country level, human rights law will be most explicitly recognised as a type of public law and will, therefore, only be enforceable against the government or state entities (if at all!).
However, domestic law regulating relations between people (i.e. civil law) can also give effect to human rights principles by giving individuals access to a remedy where their human rights have been harmed by another person or private company (e.g. data protection law, equality or non-discrimination laws). Some acts may also be criminalised in order to protect and promote the rights of others (e.g. the right to life is protected through the crime of murder).
At the same time, domestic law itself may violate international human rights law and may be challengeable before the courts on that basis (e.g. mass surveillance laws might be overturned because they violate the right to privacy).
Bringing human rights litigation involves an often-painstaking analysis of what the law requires in a given scenario and an assessment of whether and how it is enforceable in a particular case. This will require legal research, as well as legal advice.
Nonetheless, this article will briefly sketch out the starting point for figuring out where to look for sources of human rights law that may be relied on before the courts.
Primary Sources of Law
Despite the proliferation of principles, ethical codes and policy documents on artificial intelligence, these documents are not legally binding and cannot be enforced before the courts as law.
To establish what principles can be enforced before the courts, you need to first identify what is set out in primary sources of law. These are the documents that state what the law actually is. They consist of the following:
· Case Law
These are international legal agreements that are binding on those that sign and ratify (i.e. the act of giving formal consent to) them. These agreements are usually signed by states and international organisations, and constitute law between those entities, but they can also be binding on individuals and entities within those states (e.g. their citizens).
International treaties have established a legal framework for International Human Rights Law. However, some international treaties have also created other types of obligations that can have an impact on new technologies and “intelligent” machines. For example, the Vienna Convention on Road Traffic is an international treaty that is aimed at establishing universal norms on road traffic use. It was originally drafted in the 1960s and included a provision that required that every driver shall “at all times” be able to control their vehicle. This was considered to have hindered development in autonomous or self-driving cars. In 2016, an amendment came into force permitting the transfer of driving tasks to a vehicle’s automated driving technologies provided these technologies meet vehicle regulations or can be overridden/switched off by the driver.
International treaties on human rights are usually technologically neutral, meaning that rights are protected regardless of the methods or means being used to interfere with them.
Key Global Instruments
The key international human rights legal instruments have been created through the United Nations system, an international organisation founded in 1945 and made up of over 190 countries.
Civil and Political Rights
These are rights aimed at protecting the life, dignity, integrity, liberty and opinion of all people.
International Covenant on Civil and Political Rights (ICCPR): this is an international human rights treaty that sets out a number of basic civil and political rights that states are bound to protect, preserve and promote. These include rights such as the right to life and human dignity, equality, freedom of expression and assembly, the right to privacy, the right to a fair trial, and freedom from torture. It has been ratified by over 170 countries across the globe, you can see whether your country has ratified the treaty here.
Economic, Social and Cultural Rights
These are rights aimed at securing the basic social and economic conditions needed to live a life of dignity and freedom.
International Covenant on Economic Social and Cultural Rights (ICESCR): this is an international human rights treaty that sets out a number of basic economic, social and cultural rights that states are expected to work towards guaranteeing. These include the right to work, right to health, right to social security and the right to participate in cultural life. It has been ratified by over 170 countries across the globe, you can see whether your country has ratified the treaty here.
Other Relevant Treaties
International Convention on the Elimination of Racial Discrimination (ICERD): this is an international treaty that commits members to ending any “distinction, exclusion, restriction or preference based on race, colour, descent or national or ethnic origin with the purpose or effect of impairing the enjoyment on equal footing of human rights.” It has been ratified by over 180 countries across the globe, you can see whether your country has ratified the treaty here.
International Convention on the Elimination of Discrimination Against Women (CEDAW): this is an international treaty that commits members to ending any “distinction, exclusion or restriction made on the basis of sex which has the effect or purpose of impairing or nullifying the recognition, enjoyment or exercise by women” of human rights and fundamental freedoms in the political, economic, social, cultural, civil or any other field. It has been ratified by nearly 190 countries across the globe, you can see whether your country has ratified the treaty here.
Convention on the Rights of the Child (CRC): this is an international treaty that sets out the civil, political, economic, social and cultural rights that all children are entitled to. These include rights such as the right to life, the right to express their views in matters affecting them, the right to privacy and the right to education. It has been ratified by nearly 200 countries across the globe, you can see whether your country has ratified the treaty here.
International Convention on the Protection of Migrant Workers (ICMW): this is an international treaty that aims to foster respect for migrants’ human rights and establish basic principles for the protection of migrant workers and their families. It sets out the civil, political, social, economic and cultural rights to be enjoyed by migrant workers and members of their families, including rights to freedom of expression and privacy. It also establishes some more specific rights and obligations, such as the ban on collective expulsion of migrant workers and their families. It has been ratified by over 50 countries across the globe, you can see whether your country has ratified the treaty here.
Convention on the Rights of Persons with Disabilities (CRPD): this is an international treaty that sets out the civil, political, economic, social and cultural rights that persons with disabilities are entitled to. It also seeks to promote and protect the human rights of disabled people by eliminating discrimination, ensuring an inclusive education system, and protecting disabled people against exploitation, violence and abuse. It has been ratified by over 180 countries across the globe, you can see whether your country has ratified the treaty here.
Although these treaties establish rights and obligations in a technologically neutral way, and they make no reference to artificial intelligence or automated systems, they are not completely free from references to new technologies.
For example, the Convention on the Rights of the Child encourages states to use new technologies to combat disease and malnutrition, and the International Convention on the Elimination of Discrimination Against Women requires states to review and reform employment equality legislation “in light of scientific and technological knowledge.”
The Convention on the Rights of Persons with Disabilities makes reference to the need for accessibility to new technologies for people with disabilities. It obliges states to foster research and development of, and promote the availability and use of, new technologies, such as mobility aids, assistive technologies and devices, and information and communications technologies. It also requires states to remove obstacles and barriers to persons with disabilities having access, on an equal basis with others, to information and communications technologies and systems.
There is no specific international human rights treaty that deals with human rights in the context of algorithmic technologies. Furthermore, there is no specific global treaty that deals with protecting people’s data and rights in the context of automated decision-making. However, there has been an initiative to transform a regional European treaty on automatic processing of personal data (see below) into a truly “global treaty.”
Key Regional Instruments
Human rights treaties have also been adopted on a regional (Africa, Europe, Americas, Asia and Pacific Islands) basis by inter-governmental organisations. For human rights enforcement, the primary inter-governmental organisations include the African Union, the Council of Europe, the European Union and the Organization of American States.
Like the UN, they have created a number of human rights treaties. Many of these localise the international treaties outlined above. For example:
Civil and Political Rights
Charter of Fundamental Rights of the European Union (European Union, legally binding on all EU member states)
Economic, Social and Cultural Rights
Charter of Fundamental Rights of the European Union (European Union, legally binding on all EU member states)
Additional Protocol to the American Convention on Human Rights in the Area of Economic, Social and Cultural Rights (Organization of American States, see countries that have ratified it here)
Other Regional Human Rights Treaties
African Charter on the Rights and Welfare of the Child (African Union)
Council of Europe Framework Convention for the Protection of National Minorities (Council of Europe)
Inter-American Convention on the Prevention, Punishment, and Eradication of Violence against Women (Organization of American States)
Inter-American Convention on the Elimination of All Forms of Discrimination against Persons with Disabilities (Organization of American States)
These regional inter-governmental organisations have made much more headway towards establishing treaties that protect people’s data from the distinct harms that could be posed by automated systems and new technologies. The Council of Europe has drawn up a treaty that seeks to strengthen data protection in the context of automatic processing of personal information. The treaty is called the Convention for the Protection of Individuals with regard to Automatic Processing of Personal Data or Convention 108. This was the first, and remains the only, international treaty on data protection. It requires, among other things, that personal data that is processed by automated systems be obtained fairly and lawfully. It also forbids the processing of sensitive personal data, such as race, political opinions, health, religion, sexual orientation or criminal history, without proper legal safeguards.
Convention 108 has been ratified by every member state of the Council of Europe. A number of non-Council of Europe States have also become party to Convention 108, such as Argentina, Mauritius, Mexico, Morocco, Senegal, Tunisia and Uruguay. There has been a campaign for the United States of America and the United Nations to ratify or adopt Convention 108, and to make it a “global treaty.”
One of the obstacles to the creation of binding international treaties is that there is usually a threshold on the number of countries that have to ratify the treaty before it can have binding legal force. The African Union has drawn up a data protection treaty, the African Union Convention on Cyber Security and Personal Data Protection, but it has only been adopted by eight states out of the fifteen ratifications needed for it to come into force.
The Council of Europe has also taken steps to modernise Convention 108, with the inclusion of stronger protections against automated decision-making systems. This is referred to as Convention 108+. It has been ratified by ten member states of the Council of Europe, well below the number required for it to come into force.
Even though there is a lack of specific global treaties that deal with the human rights implications of “intelligent” machines and other new technologies specifically, it has not stopped the enforcement of the technologically neutral rights contained in international treaties in cases concerning such systems.
For example, the European Convention on Human Rights has been successfully relied on to challenge the use of live automated facial recognition technology and a risk scoring model that was used to identify those at risk of committing welfare fraud.
This is (usually) a written document that sets out the basic structure of a state, the major state institutions (such as the government, legislator and the courts), and the principles governing their relations with each other and with citizens. It is, therefore, often treated as the fundamental and supreme law of a country.
Constitutions across the globe include human (or constitutional) rights that can be enforced against the state and state bodies, such as the rights to privacy, freedom of expression, a fair trial and freedom from discrimination.
Some constitutional rights may not be explicitly referred to as “human rights,” but can play a similar function to the rights set out under international treaties. For example, in the US, constitutional rights to due process can offer similar protections to the right to a fair trial. As set out in two case studies, due process rights have been relied on to challenge the use of risk scoring algorithms in the criminal justice system and the use of student performance scoring algorithms used to fire teachers in public education contexts.
Constitutions may directly incorporate international treaties into the domestic legal framework of a country, allowing individuals to enforce their international human rights against state bodies. For example, the Constitution of Kenya states that “the general rules of international law shall form part of the law of Kenya.”
Constitutions may require that domestic laws be interpreted in conformity with international treaties. For example, the South African Constitution states that courts and tribunals must consider international law when interpreting constitutional rights.
However, other countries require that, before citizens can enforce their rights under international treaties before the courts, domestic law has to be passed to make international law directly applicable.
Legislation or Code
These are the laws that are passed (or enacted) by the legislature. The legislature being the deliberative assembly with the authority to make laws in a given country.
Legislation sets out legal norms that can apply in a variety of contexts (see above on criminal, civil and public law). Some countries have legislation that incorporates international human rights law into the domestic legal context, permitting these rights to be enforced before the courts (e.g. the Human Rights Act in the UK). Legislation may elaborate, specify or complement further what has been set out in international human rights treaties (e.g. domestic data protection legislation has done this).
Some legislation may narrowly apply to the design, development or use of machine learning and automated decision-making technologies specifically. For example, Canada has passed a law that specifically regulates government use of automated systems in circumstances where public services are being provided and where recommendations or decisions are being made about a person to a government department.
A legal code is a type of legislation that is more prevalent in civil law jurisdictions. It attempts to exhaustively and comprehensively set out a written statement of laws that cover the entire legal system or a particular area of law (e.g. the French Code Civil).
Legislation can also be further categorised into primary and secondary legislation. Primary legislation is the main set of laws passed by the legislature. Secondary legislation refers to laws that are made by lesser bodies that are granted powers to issue delegated legislation.
Secondary legislation or policies can usually be challenged for being outside the powers authorised under primary legislation. This was done successfully in a case before the Constitutional Tribunal in Poland, for example, where the fact that the Minister of Labour was able to put in place a process of algorithmically profiling unemployed individuals was beyond the powers that could be delegated to such an individual under primary law.
The constitution, primary legislation and secondary legislation function in a hierarchy, which determines how they rank in authority. The structure of this hierarchy can vary from country to country. However, it can be possible to challenge legislation for its failure to comply with the constitution and, in some countries, international human rights law.
As a result, a piece of legislation that requires internet platforms to adopt automated content moderation technologies or that regulates the circumstances under which facial recognition technology can be used, could still be challenged before the courts for failing to sufficiently protect the rights to freedom of expression, privacy or freedom from discrimination.
Legislation can be amended or repealed by later legislation and may even be overturned by the courts. Therefore, when researching domestic legislation, it is always worth doing some further digging to make sure it is still in force and unaffected by later developments.
Numerous legal systems recognise the binding authority of court decisions as law. The body of law made up of case law is usually referred to as judicial precedent.
Under these systems, courts will be bound to follow previously decided cases from courts of superior authority when they are presented with a case of sufficient similarity to the facts of previous cases.
When deciding a case, courts may consider previous case law in a number of ways:
Approve: the court states that the decision of an inferior court was correctly decided.
Apply: the court applies the reasoning of a previous decision to the case before it.
Follow: the court is bound to follow the reasoning of a superior court where the facts are substantially the same.
Distinguish: the court does not follow an otherwise binding decision because there is a relevant difference in facts.
Disapprove: the court might state that the decision of a lower court in a different case was wrongly decided.
Doubt: the court might give reasons why a court decision in another matter may have been wrongly decided.
Not Follow: the court might refuse to follow the decision of a court of equal status even though the facts of the case are substantially the same.
Overrule: the court may expressly overrule the decision of an inferior court in another case.
As with legislation, case law can evolve over time. For example, case law may be superseded by legislation. Court decisions can also be appealed to higher courts. This might result in a higher court approving or applying a lower court decision, or it could result in the lower decision being overruled (meaning it will no longer be good law). This means it is crucial to make sure that a court decision has not been rendered invalid or of limited value by other case law or legislation before it is relied on to take a case to court.
The reasoning of a court decision can, itself, be separated into different levels of authoritative legal weight. You can find out more on this here.
Case law will only be binding within a specific country or legal system. Nonetheless, case law from courts in other countries on similar or the same matter can be of persuasive value to some other courts. For example, a court in France might take account of how the courts in Germany dealt with a similar legal question.
Not (Really) Law
If a principle is not set out in a primary source of law, it will not be legally binding as law. Meaning it will not be enforceable before the courts.
For example, a range of ethics-type documents have been drawn up in relation to artificial intelligence. These have been published by companies (e.g. IBM, Google, Microsoft, ITI, Telia Company), international organisations (e.g. European Union, Organisation for Economic Co-operation and Development, G20), and other stakeholders (e.g. AI Industry Alliance, New York Times, IEEE, Future of Life Institute, Partnership on AI, University of Montreal, and the Beijing Academy of AI). More examples can be found on Algorithm Watch’s AI Ethics Guidelines Global Inventory.
These ethical documents are aimed at guiding the development of artificial intelligence toward morally and socially desirable ends. Although courts may review them when considering cases, they will not be bound to apply or follow them since they do not carry the force of law.
Similarly, a number of documents have been published by governments across the globe, such as in Belgium, China, Czechia, Denmark, France, Germany, Hungary, India, Italy, Japan, Lithuania, Luxembourg, Mexico, Saudi Arabia, Serbia, Sweden, and Norway, that set out a national strategy on artificial intelligence. These also do not constitute law, although courts may refer to them when trying to understand the purpose and intention behind a law that has been adopted pursuant to these strategies.
Materials that constitute secondary sources of law, i.e. materials that discuss and comment on the law, are also not legally binding as law. However, courts may consider these materials when trying to interpret and apply the law. These might include textbooks, legal dictionaries, encyclopaedias and journal articles.
There are a number of agreements, guidelines, principles and declarations that can be considered soft law. Again, these are not legally binding but may influence how laws are interpreted and applied by the courts. Some that are most relevant to the interpretation and application of international human rights law to machine learning and similar technologies are referred to below:
General recommendation No. 36 (2020) on preventing and combating racial profiling by law enforcement officials. This is a recommendation drafted by the Committee on the Elimination of Racial Discrimination that provides guidance on preventing and combatting racial profiling in law enforcement. The recommendation includes agreed definitions of racial profiling, and the relevant human rights principles and obligations owed under international law. The document includes specific recommendations on artificial intelligence, including a call on states to investigate and impose sanctions against all instances of algorithmic bias.
Reports of the Special Rapporteur on contemporary forms of racism and related intolerance, one that looks at different forms of racial discrimination in the design and use of emerging technologies (2020) and another that looks more specifically at the discriminatory impact of emerging digital technologies on migrants, stateless persons, refugees and other non-citizens (2020). These reports highlight the threat posed by technologies, including artificial intelligence, through the perpetuation and entrenchment of direct, indirect and structural discrimination. The second report outlines how machine learning and other technologies are used to put up digital borders, and in doing so acknowledges the historical antecedents to these technologies in colonial technologies of racialised governance.
Report of the Special Rapporteur on extreme poverty and human rights that looked at the Digital Welfare State (2019). This report highlighted the increased reliance by governments across the world on digital and algorithmic systems in their provision of social security or welfare to individuals. It recommends that government technologies be directed towards ensuring people a higher standard of living rather than cost savings, efficiency and penalising individuals.
Reports of the Special Rapporteur on the right to freedom of opinion and expression on artificial intelligence technologies and implications for the information environment (2018), and another on online content moderation (2018). The first of these reports looks at how algorithms and artificial intelligence technologies are pervading the information eco-system, from digital devices to search engines and social media platforms. The report proposes a human rights framework for the design and use of such technologies. In the latter report, the Special Rapporteur takes a specific look at the human rights implications of machine learning and other automated technologies in content moderation. He noted that, in order for individuals to enforce their right to freedom of expression where automated systems have been used, people must know that they have been subject to an algorithmic decision and they should be equipped with information about the logic behind the decision.
Report of the UN Independent Expert on the enjoyment of all human rights by older persons (2017). This report provides an analysis of international human rights law protecting the rights of older persons and applies the relevant principles to the adoption of robots and artificial intelligence in the care of older persons. While recognising the benefits of technological advancements, the Independent Expert highlighted threats posed in areas such as accessibility, privacy, health, safety, liberty and non-discrimination.
Report of the United Nations Commissioner for Human Rights (2017). This report looked at the threats posed to human rights by the gender digital divide (the gap that exists between genders in their access to, use of, and ability to influence, contribute and benefit from information technologies). He highlights that gender equality should be promoted in the design and implementation of information technologies, and in the frameworks to regulate them. He also stated that, in relation to “artificial intelligence,” efforts should be made to ensure that data inputs are inclusive and accurate and that they operate in a human rights compliant way. He concluded “[w]ith regard to new and emerging data-driven technologies, there is now a critical window of opportunity to ensure that these technologies are human rights compliant and do not replicate or exacerbate existing patterns of discrimination against women.”
Report of the Special Rapporteur on extrajudicial, summary or arbitrary executions on Killer Robots (2013). This is a report by a UN Special Rapporteur that outlines the human rights implications of “killer robots” or Lethal Autonomous Weapons. Although he did not reach any hard conclusions on the extent to which these weapons can comply with human rights law, he recommended that the introduction of such technologies be treated with extreme caution. He recommended a moratorium on such technologies until their permitted development and use could be ascertained.
The United Nations Guiding Principles on Business and Human Rights. This is a document that sets out guidelines for states and businesses on how to prevent, address and remedy corporate human rights abuses.
Council of Europe
Recommendation of the Committee of Ministers to member States on the human rights impacts of algorithmic systems - CM/Rec(2020)1. This document adopts specific guidelines for states to meet their obligations to protect and promote human rights and fundamental freedoms in the context of algorithmic systems. It also sets out the responsibilities of private actors with respect to human rights in this context.
Recommendation on developing and promoting digital citizenship education - CM/Rec(2019)17. This document includes a set of guidelines around the education of citizens so they can participate fully in the affairs of their community as digital citizens. The guidelines include specific principles around the use of machine and deep learning in the provision of education (e.g. in providing predictions of teaching patterns and learning outcomes). The guidelines state that it is necessary to take human rights principles into account in the early design of the application of such technology. The guidelines warn against the excessive profiling of learnings, the overuse of analytics that can over-standardise educational goals and learning outcomes, potential discrimination and exclusion bias, and infringement of privacy rights posed by such tools.
Recommendation of the Committee of Ministers to member States on preventing and combating sexism - CM/Rec(2019)1. This document includes guidelines for preventing and combating sexism, including in online or digital environments. The guidelines note the role that algorithms can play in transmitting and strengthening gender stereotypes. The guidelines call on states to “[i]ntegrate a gender equality perspective in all policies, programmes and research in relation to artificial intelligence to avoid the potential risks of technology perpetuating sexism and gender stereotypes.” It also asks states to examine how such technology could help close gender gaps and eliminate sexism.
Declaration of the Committee of Ministers on the manipulative capabilities of algorithmic processes - Decl(13/02/2019)1. This document draws attention to the risks involved in using algorithmic systems to manipulate social and political behaviour (e.g. through micro-targeting techniques). It warns that this manipulation can hinder individuals from formulating their opinions and taking decisions independently of automated systems. The Declaration asks states to consider the “need for additional protective frameworks related to data that go beyond current notions of personal data protection and privacy and address the significant impacts of the targeted use of data on societies and on the exercise of human rights more broadly”.
Guidelines on Artificial Intelligence and Data Protection - T-PD(2019)01. This is a document drafted by the Consultative Committee of Convention 108. It provides a set of baseline measures that governments, developers, manufacturers, and service providers should follow to ensure that artificial intelligence applications do not undermine the human dignity and the human rights and fundamental freedoms of every individual, with a particular emphasis on data protection
Unboxing AI: 10 steps to protect human rights, recommendation of the Commissioner for Human Rights (2019). This document sets out a number of recommendations for countries, as well as those involved in the development, implementation or impact of artificial intelligence, in order to help ensure effective observance and full enjoyment of human rights in the context where such technologies are being used. It includes recommendations on the carrying out of human rights impact assessments, public consultations and independent oversight in relation to the development, design or use of artificial intelligence. It also has recommendations on how specific rights can be safeguarded from these technologies.
European Ethical Charter on the use of artificial intelligence (AI) in judicial systems and their environment - CEPEJ(2018)14. This is a non-binding charter that sets out a framework of principles related to the use of automated technologies in the justice system. There are five core principles, including the principle of respect for fundamental rights, non-discrimination, quality and security, user control, and transparency, impartiality and fairness.
Recommendation of the Parliamentary Assembly of the Council of Europe about Technological convergence, artificial intelligence and human rights – Recommendation 2102(2017). This document calls for the drawing up of guidelines that strengthen transparency, regulation, and accountability in relation to automated systems. It also calls for the establishment of standards to be complied with when courts use artificial intelligence. Furthermore, it asks for “the recognition of new rights in terms of respect for private and family life, the ability to refuse to be subjected to profiling, to have one’s location tracked, to be manipulated or influenced by a ‘coach’ and the right to have the opportunity, in the context of care and assistance provided to elderly people and people with disabilities, to choose to have contact with a human being rather than a robot.”
Recommendation CM/Rec(2010)13 adopted by the Committee of Ministers of the Council of Europe on the protection of individuals with regard to automatic processing of personal data in the context of profiling. This document establishes a set of recommendations to protect personal data in contexts where it is used to profile them as having certain characteristics. It includes recommendations on the conditions under which such data may be processed, the transparency of such processing, and the rights of individuals whose data is being used.
EU soft law includes quasi-legal instruments such as opinions, recommendations, communications, non-legislative resolutions, notices, and guidance documents.
The European Parliament has initiated non-binding resolutions that provide interpretations of human rights law in relation to machine learning technologies. For example, in the EU legislative Resolution on a Comprehensive European industrial policy on artificial intelligence and robotics, the European Parliament expressed “great concern about the employment of AI applications, including facial and voice recognition, in ‘emotional surveillance’ programmes, i.e. monitoring the mental conditions of workers and citizens in order to increase productivity and preserve social stability, sometimes coupled with ‘social credit’ systems, as already seen in China, for instance” and stressed “that such programmes are inherently at odds with European values and norms protecting the rights and freedoms of individuals.” (see also Resolution on Civil Law Rules on Robotics and AI).
The European Data Protection Board, a body responsible for ensuring consistent application of data protection laws across the EU, has adopted guidelines on the application of such laws to automated individual decision-making and profiling, as well as connected vehicles and mobility aids.
These are where to find the fundamental principles of human rights law, but what are the rights themselves and how to they relate to machine learning technologies? This is explained further in the explainer on Spotting There Is A Problem.
If you notice anything incorrect or missing from our explanations, please let us know through the contact form!