Working Out A Strategy


There are many reasons why an individual or an entity may bring a case to court. From trying to settle a legal disagreement with another person, to trying to bring about some kind of systemic change to our society.


The vast majority of cases that are brought before the courts are for individual redress. In other words, an individual will be seeking to resolve a legal wrong that has been committed against them.


However, some cases will go further and seek a change beyond the immediate interests of the person taking the case, such as a change to law, practice or behaviour. This is often referred to as strategic litigation. There are a variety of freely available resources that explain and provide guidance on strategic litigation (including here, here, here, here and here).


Regardless of the nature of the case being brought, it is always worth devising a strategy with a lawyer before taking it.


A legal strategy will seek to identify the key action points and considerations that need to be taken in order to maximise the potential of achieving the outcome sought from the case and minimise the risk for those involved in pursuing it.


It may just be that, after developing and reflecting on the strategy, it is decided that litigation is not an appropriate course of action.


There are no hard rules on what should be included in a legal strategy, but various factors that may be considered include:

This article will sketch out some of the issues that might form part of the thinking in answering each of these questions.



What solution is needed to obtain justice?


When deciding whether to take a case, the first thing that should be identified is what the goal is in pursuing the litigation.


This goal should preferably be something that is specific, measurable and realistic. For example, it may be to stop the use of a particular technology by a particular entity or it may be to get financial compensation for specific individuals harmed by a particular system.


When someone takes a case to court, they are effectively asking the court for a solution to the problem brought before it. This solution is called a remedy. Remedies are orders made by the court that are designed to make amends for the legal wrong that has taken place.


One goal in taking a case may be to obtain a specific kind of remedy that the court can provide. Alternatively, it may be that the goal transcends or is unrelated to a particular remedy and can be achieved purely through the action of taking litigation on a particular issue.


When taking litigation, it is important that an assessment is carried out on what achievement is sought and to what extent litigation can help towards that goal.


There are a number of goals that can be obtained through the taking of legal action to court, some are considered below:

  • Court recognition of a position: one remedy available to courts is to provide a declaration. This is an authoritative statement in support of a particular position. For example, the court might declare that the rights of an individual had been violated or that use of a certain technology in a particular context was unlawful. A declaration does not involve the court explicitly telling someone to do something, but it can offer vindication that can then be used to push for a certain change. Many of the cases that have concerned automated systems to date have concluded with the court making a declaration. In Ewert v. Canada, for example, the Supreme Court of Canada declared that an authority responsible for managing prisoners had broken the law in its use of actuarial tools to assess an inmate’s risk.

  • Law reform: a decision from a court can push the government and legislature to abolish or amend a particular law. Some courts may even have the power to invalidate or overturn laws themselves, or they may order or declare that the government take action to reform the law. In the SyRI case, for example, a court in the Netherlands invalidated a law that established a system of using risk scoring models in the detection of welfare and other types of fraud.

  • Clarity on the law: a decision will involve the interpretation and application of law to a specific set of facts. In this way, courts can provide clarity on how a law, e.g. a human right, should be interpreted or applied in certain contexts. Such decisions can form part of the case law that must then be followed in future cases, particularly when the future cases are decided by lower courts (what is meant by "lower courts" is explained here). This clarity might also compel other individuals or entities to change their future behaviour because they can no longer plead uncertainty around what is required of them under the law. In Loomis v. Wisconsin, for example, the Supreme Court of Wisconsin, while interpreting and applying constitutional law, clarified the limited set of permissible circumstances under which a court could refer to an algorithmically-produced risk score when sentencing people. This was despite the fact that the court ultimately upheld the use of such a score in the individual case before it.

  • Establishing rights: some courts may establish or recognise rights that were previously unavailable, unrecognised or unclear, giving individuals more freedoms or entitlements than had previously been thought to be the case. In a recent landmark decision in the US, for example, a court found that researchers testing whether algorithms on employment sites were discriminatory by creating false profiles and otherwise breaching websites’ terms of service were not breaking laws against computer fraud and abuse. This cleared the way for academics, computer scientists and journalists to carry out this research without fearing that they were breaking the law.

  • Making amends: courts are tasked with remedying or redressing the legal harm that is brought before them. In this way, they try to rectify certain situations and put a person who has been harmed as close to the position they would have been in had the harm not occurred. Courts often do this by ordering one party to pay damages, or compensation, to the party that was harmed. Damages can relate to economic loss, as well as physical or psychological harm, sustained by those harmed. In Italian General Confederation of Labour v. Deliveroo, an Italian court ordered Deliveroo to pay €50,000 to a trade union in compensation for the harm caused to Deliveroo riders by the systemic use of a discriminatory algorithm.

  • Order to do (or not do) something: courts can also compel an individual or entity to take a certain course of action or refrain from doing something that will potentially result in a legal harm. This might also involve the court stopping something from happening again. These court orders are usually referred to as injunctions. For example, the court may forbid the further use of a particular technology or system. There was a recent example, although not strictly speaking from a court, of an app developer being required to delete all data, models and algorithms produced from unlawfully obtained data.

  • Fact-finding: even if unsuccessful, legal cases can shed light on information previously kept secret. Court procedures often involve the compulsory disclosure by each party of relevant documents, which can bring more transparency to how an automated system has been designed, procured, deployed or used. Furthermore, by bringing another party before a court, it is possible to scrutinise and question the party’s explanation for why and how they use a certain type of technology. In Houston Federation of Teachers and Others v. HISD, for example, a school district ended up admitting before a court that they did not take steps to verify or audit a secret algorithm that they had been using to dismiss teachers for poor performance.

  • Putting in place safeguards: a court might not completely outlaw the use of a certain technology. Instead, it might deliver a decision requiring that certain safeguards be put in place in the design, deployment or use of such technology in order to protect human rights. This will ensure that, if such technology is eventually used, there is minimal risk that it will cause a violation of individual rights. In R (Edward Bridges) v. South Wales Police, for example, the court’s decision gave guidance on the kind of law that needs to be in place, and the level of assessment that needs to be carried out, in relation to facial recognition technology before it may be deployed by police.

  • Raising awareness: taking a case before the courts can also raise public awareness of certain issues. Litigation can be accompanied by increased publicity and media attention on the legal problem before the court. This may, for example, bring home to more people the human rights implications of certain technologies. The court itself may provide remedies that help raise the issue in public consciousness, and even shame the responsible party for their unlawful conduct. In Italian General Confederation of Labour v. Deliveroo, for example, the court ordered that Deliveroo publicise on their website and in a national newspaper that a court had found an algorithm they had used to score workers to be discriminatory.

  • Causing fuss: as noted below, taking a case can be expensive, time consuming and risky, but so can defending a case. For some individuals or entities, being brought before a court to account for something they have done or failed to do will be enough to convince them to change their ways. As a result, sometimes simply the threat of bringing a case can produce a positive outcome for human rights. There have been a few examples of this. In the Deliveroo case mentioned above, Deliveroo got rid of the algorithm being challenged before the court shortly after the case was filed (most likely in the hope that this would stop the case in its tracks, which it didn't). In the UK, a digital rights organisation, Foxglove, was able to get the government to scrap an algorithm used to grade student exams and an algorithm used to process visa applications after threatening to take them to court.

Legal strategies may include more than one of these goals.


The strategy may even be to pursue these goals through multiple legal cases in order to (eventually) bring about the change that is ultimately sought.


For example, litigation may initially be brought to increase public debate around the harmful use of a certain technology, followed by a case to get clarity on what the law says about how that technology should or should not be used, before finally obtaining a court decision that forbids the use of the technology or overturns the law permitting its use.



Is there a legal route to this solution?


Another factor worth considering in a legal strategy is the extent to which a legal route exists to obtaining the goal that is sought. In other words, is there a legal remedy available that can achieve (or help someone achieve) the outcome they are looking for?


To give a somewhat reductive example, trying to get a court to enforce a law that permits or requires the use of a risk scoring algorithm will be unlikely to stop the use of that algorithm in the future. Instead, this would be more likely to be achieved if the law permitting or requiring the use could be challenged on the basis that it violates the right to privacy or the right to a fair trial.


The best or most appropriate legal route to pursue can be tricky to work out, and requires a sound understanding of what legal options are available in a given scenario.


This means having a comprehensive overview of:


What forum is best suited to offer the solution?


Life would be so much easier if there was only one court where we could bring all our legal cases to.


In reality, and for good reason, there are a variety of courts that have different areas of competence and different levels of authority.


These courts might also only be able to provide certain kinds of remedy in the cases that have been brought before them. For example, not all courts can overturn a law.


Therefore, as part of a strategy, a person taking a case will have to consider what forum is best suited to achieving the outcome they want to achieve.


As well as considering what a relevant court can or cannot do, this aspect of a strategy also involves consideration of a range of other factors.


For example, it might involve consideration of whether a court has the technical expertise needed to make the correct findings in a case. It may involve looking at how the court has handled cases involving similar issues in the past, to assess the likelihood of them providing an interpretation and application of the law that will provide the desired outcome. It could include a consideration of the extent to which governments or other entities comply with the decisions of the court (as noted here, some court decisions are not complied with or are difficult to enforce).



Who should take the case and what role do others play?


It may be that some individuals or entities are better placed to take a case than others. For example, maybe only certain people will be allowed to bring the claim to court (see here for more information on standing, which determines who can bring a case before certain courts), or some people might have a stronger case than others to allege that their rights have been violated.


Therefore, for those thinking of taking legal action, it is worth considering whether they are the right person or entity to take the case at the relevant time.


If they are not, the question they might try to answer is “what can we do to support and maximise the chances of success for those who are taking such a case?”


There are many ways in which individuals and entities can engage in litigation (without being the ones taking the case), for example:


  • Represent: if you are a lawyer or a legal charity with the capacity and resources to take on a case, you might decide to represent the person pursuing the legal claim. This may be by acting as their lawyer, or it may be by taking action to court on their behalf (something called representative action). In many countries, there are strict limits around how an individual or organisation can represent others before a court, so it is worth looking into what you can or cannot do if you decide to represent an individual in their legal action.

  • Provide Input: it may be more appropriate to simply share information and resources with those taking the case. This may be by providing legal input, helping them understand the technology they are seeking to challenge, sharing evidence or giving them access to materials you have used in similar cases, without officially representing them before the court.

  • Campaign and Publicise: it is possible to support litigation by publicising and raising public awareness of the issues at stake in the litigation. This could help in achieving the objective of the case, for example, if the case has been taken with the purpose of raising the issue in public consciousness. It can also help put pressure on the individual or entity against whom the case has been taken to do the right and just thing in the circumstances of the case. If you have unique or special access to decisionmakers who can bring about the solution sought through the case, you may offer to engage with those decisionmakers on the issues raised in the litigation.

  • Intervene: some courts can also accept the input of third parties. This means that an individual or organisation can provide information directly to the court so that it can reach the correct decision according to the law. This input may be in the form of factual evidence or legal analysis. There have already been some examples of third-party interventions (sometimes referred to as amicus briefs) in cases concerning automated systems or machine learning. For example, in the SyRI Case, the UN Special Rapporteur on extreme poverty and human rights filed a third-party intervention highlighting that the statistical model that was used to help detect welfare fraud in the Netherlands had a discriminatory and stigmatising effect in violation of international law. In Ewert v. Canada, a host of organisations provided third-party interventions on the discriminatory impact of risk scoring models in the criminal justice system, including the Native Women’s Association of Canada, the Mental Health Legal Committee, the Canadian Human Rights Commission, Aboriginal Legal Services, and the Union of British Columbia Indian Chiefs.

It may be that a person or entity that has taken a legal case will still want to support and assist in litigation efforts that complement or pursue similar goals to their own. In such circumstances, they might also consider taking on the above roles in relation to those other cases.



Who should the case be against?


A case must be taken against an individual or legal entity (such as a company or government body) responsible for the relevant legal harm alleged.


This requires that those taking the case show that they have an enforceable right against this particular person or entity.


This can be difficult to figure out and prove, particularly when complex and hidden structures and lines of accountability are at play. Nonetheless, answering this question usually involves a process of:

  • identifying who is bound by a relevant law in a particular circumstance;

  • assessing whether they have broken that law; and

  • determining whether a court has competence to decide a case in relation to that person or entity.

It may be that a person or entity involved in developing or designing a technology, for example, is not bound by the law on which a case rests. This would mean that they cannot be taken to court pursuant to that law. For instance, a law that regulates when a public authority can use certain kinds of technology will not be enforceable against a private company that designs the technology itself.


In some circumstances, a person taking a case might have a range of potential individuals or entities that they have a valid legal claim against. In which case, other factors might have to be considered as part of a legal strategy. For example:

  • Should cases be taken against all persons or entities at the same time, or should there be staggered or selective legal claims?

  • Which individual or entity is there the strongest case against (e.g. there might be more evidence of one individual’s or entity’s unlawful conduct over others)?

  • Does bringing a case against one individual or entity bring you closer to your objective for pursuing litigation (e.g. are they the most likely to bring about the solution you are seeking. An entity with few financial resources, for example, might not be able to compensate the person taking a case)?

  • Would a case against one person or entity be more cost-effective and less risky (see