Review into Bias in Algorithmic Decision-making
United Kingdom · Guideline · Centre for Data Ethics and Innovation
Our summary
This report examines how algorithmic tools affect bias in high-stakes decisions across recruitment, finance, policing and local government. It proposes that organisations should proactively measure and manage algorithmic bias using data while maintaining human oversight for life-affecting choices. The review recommends that regulators, industry bodies and government develop sector-specific standards to ensure algorithms promote fairness rather than entrench discrimination.
Written by the Index in our own words. It is not legal advice; the official text is the authority.
Inside this policy
Every section of this policy in order. Our notes summarise each one in plain language and do not quote the official text, which is the authority.
Drafted by the Index An automated summary of the official text, not yet reviewed by a person.Reviewed Checked by a person against the official text.
87 sections
Jump to a chapter
- Preface
- Executive summary
- 1 Background and scope
- 2 The issue
- Part II: Sector reviews
- 3 Recruitment
- 4 Financial services
- 5 Policing
- 6 Local government
- Part III: Addressing the challenges
- 7 Enabling fair innovation
- 8 The regulatory environment
- 9 Transparency in the public sector
- 10 Next steps and future challenges
- Appendix A
- Is this page useful?
Preface
Automated and data-driven approaches can reveal existing biases but risk entrenching or creating new ones, and organisations particularly in the public sector should design processes allowing human judgment to intervene and achieve fair outcomes. Organisations have an ethical obligation to use data to identify and mitigate bias in policing, finance, recruitment and social services, with regulators and industry bodies establishing sector-specific standards for anticipating bias, auditing algorithms and addressing problems through skilled professionals. Definitions of fairness in algorithmic systems are diverse and sometimes contradictory, requiring society-wide agreement, and organisational leaders must remain publicly accountable for decisions regardless of whether algorithms or humans make them, with transparency about how algorithms are used and what checks exist.
Executive summary
2 sections
The review examines how increasing use of algorithms affects bias in decision-making across four sectors, recognising that algorithms can encode human biases but may also improve fairness through better data use if applied carefully. It concludes that algorithms cannot handle all aspects of decision-making and that human judgement is essential in some contexts.
Sector reviews
Algorithmic decision-making tools are expanding across recruitment, financial services, policing and local government but lack adequate governance frameworks to prevent discrimination and bias. Recruitment tools need clearer guidance on preventing discrimination with historical data; financial services systems risk entrenching historic biases in credit decisions; policing requires national coordination to balance security, privacy and fairness with democratic oversight; and local government's predictive technologies risk overrepresenting certain groups and producing disproportionately high error rates for minority groups. The report recommends that the Equality and Human Rights Commission revise guidance on the Equality Act 2010 for recruitment involving algorithms, the Information Commissioner's Office collaborate with industry to improve consistent application of guidance, the Home Office clarify roles and responsibilities for national policing bodies on data analytics, and government establish national guidance for local authorities on legally and ethically procuring or developing algorithmic decision-making tools for significant decisions about individuals.
Addressing the challenges
The report identifies three main areas for creating fair algorithmic decision-making: providing organisations with necessary tools and knowledge, establishing regulatory frameworks that incentivise fair practices, and encouraging public sector transparency and leadership. These elements are interdependent, requiring both support for ethical innovation and clear accountability standards.
Enabling fair innovation
The regulatory environment
Public sector transparency
Next steps and future challenges
1 Background and scope
4 sections
1.1 About CDEI
The Centre for Data Ethics and Innovation is an independent committee of experts established by the UK government to study and provide advice on maximizing benefits from data-driven technologies. Its objective is to foster an environment where ethical innovation flourishes, public confidence in technology development is maintained, algorithmic decisions are trustworthy, and innovation risks are managed.
1.2 About this review
The government commissioned CDEI in October 2018 to investigate potential bias in algorithmic decisions. This final report includes formal recommendations to government and provides advice for regulators and industry on responsible innovation and trustworthy governance.
1.3 Our focus
The review prioritises algorithms making or informing decisions with considerable impact on individuals, such as those meeting Data Protection Act significance thresholds. It focuses on current or near-future algorithmic use across sectors and examines biased outcomes rather than other forms of unfairness, whilst acknowledging that smaller decisions with cumulative bias effects fall outside the core scope.
1.4 Our approach
Sector approach
Cross-sector themes
Evidence
2 The issue
6 sections
This section outlines the growth of algorithmic decision-making across society and the accompanying concerns about unfair outcomes caused by systematic skew. It establishes that bias can occur through reinforcing past prejudices or creating new ones, distinguishes between bias that constitutes unlawful discrimination and other forms of unfair algorithmic bias, and explains that fairness extends beyond removing bias to include being non-arbitrary, reasonable, and respectful of individuals' circumstances.
2.2 The use of algorithms in decision-making
Algorithmic decision-making encompasses any choice procedure in which an algorithm either makes or substantially guides the outcome, including cases where a human retains final authority, and is expanding across sectors due to increased data availability, cheaper computing power, cost-saving pressures, and advances in machine learning techniques. The review focuses mainly on machine learning approaches applied to decisions affecting individuals, though much of the content relates to other rule-based procedures. Algorithms typically form just one component of a broader decision procedure alongside human judgment, exceptions, and avenues for reconsideration or appeal, and achieving fair outcomes requires examining the entire process rather than only the algorithmic component, with human involvement ranging from substantial oversight to complete absence in fully automated systems.
2.3 Bias
Bias in algorithmic decision-making refers to outputs that are not only skewed but unfairly skewed, and can arise through historical bias in training data reflecting past inequalities, data selection bias making data unrepresentative of certain groups, algorithmic design bias such as optimising engagement at cost, human overseer bias re-entering the process, and feedback loops that amplify bias over time as models retrain on data partly generated by earlier model versions.
2.4 Discrimination and equality
This section applies the legal definition of discrimination from the Equality Act 2010, meaning unfavourable treatment based on protected characteristics such as age, race, sex, or disability in public functions and service provision. It distinguishes direct discrimination, where someone is treated less favourably because of a protected characteristic, from indirect discrimination, where a policy disadvantages a group sharing a protected characteristic without legitimate reason.
Unfair bias beyond discrimination
2.5 Fairness
Overview
Notions of fairness
Addressing fairness
Why algorithms are different
2.6 Applying ethical principles
Bias elimination in algorithmic and human decision-making is context-dependent and unlikely to be complete. Organizations should apply five OECD principles to AI systems: benefiting people and the planet, respecting rule of law, human rights and democratic values with appropriate safeguards, ensuring transparency for understanding and challenging outcomes, functioning robustly, securely and safely throughout their lifecycle, and ensuring accountability for proper functioning. When introducing algorithms into decisions previously handled only by humans, organizations should attain standards of openness, responsibility and impartiality comparable to those applied to human decision-making and often exceed them, including robustness and security throughout the system lifecycle, clear accountability lines, and structure and procedures to detect, manage and remedy potential unfairness by making explicit the value judgements and trade-offs embedded in algorithmic rules.
2.7 The opportunity
Recommends that organisations using algorithms to make significant decisions should measure bias through data, address any historical patterns found, design bias mitigation when replacing human systems, and assess fairness by looking at whole process impact rather than algorithm outputs alone. Humans in supervisory roles need sufficient understanding of how models work to make informed judgements about their effectiveness and fairness.
Part II: Sector reviews
This part examines bias issues in four sectors: recruitment, financial services, policing and local government. Each sector involves large-scale decisions affecting individuals, shows growing use of algorithmic tools, and has historical evidence of bias that algorithms might perpetuate. The sector-specific findings inform broader recommendations.
3 Recruitment
Algorithmic decision-making in recruitment is increasingly used in sourcing, targeting advertisements, CV screening, video interviewing and selection, but training algorithms on historical data risks embedding bias from patterns lacking relevance to job performance or from undiverse datasets; organisations must rigorously test models to identify proxies correlating with protected characteristics and avoid indirect discrimination, monitor outcomes across individuals with overlapping protected characteristics, and build inclusive design with human-assessed alternatives for disabled people. Recruitment tool providers are largely self-regulated with internal testing, though the Equality and Human Rights Commission provides no specific guidance on Equality Act 2010 application and researchers call for greater transparency and independent auditing; collecting demographic data is standard practice for verifying direct or indirect discrimination, though the US four-fifths rule lacks UK application. Fully automated rejections are regulated under Article 22 GDPR, requiring organisations to inform individuals about processing, enable requests for human intervention or challenge, and conduct regular system checks, though research suggests this guidance is rarely applied consistently in practice, particularly regarding human review provisions.
4 Financial services
1 section
Financial services organisations vary in their sophistication and risk tolerance with algorithms, and hold mixed views on collecting data about protected characteristics needed to detect bias. The review notes that clearer model explainability is important for identifying discriminatory outcomes and building customer trust, and that the Financial Conduct Authority provides stronger regulatory oversight than other sectors studied.
Current landscape
5 Policing
1 section
Algorithmic adoption in UK policing is still emerging, with variable approaches across forces. Risks include systematically unfair outcomes and tension between security, privacy and fairness objectives. The review recommends that the Home Office establish clearer national leadership and responsibility for data ethics standards in policing, as current governance is fragmented across multiple bodies.
Our approach
6 Local government
1 section
Local authorities increasingly use data-driven tools for service decisions, though most remain early-stage. The review identifies data quality and infrastructure as major barriers to deployment and recommends that government develop national guidance to help local authorities ethically procure or develop these tools, including specific advice on identifying and reducing bias.
Tool development approaches
Part III: Addressing the challenges
Part III outlines three main areas for addressing bias in algorithmic decision-making: enablers to help organisations build fair systems, regulatory incentives to encourage fairness, and public sector leadership through transparency. It assesses progress across six dimensions (algorithm design, bias measurement, bias mitigation, expert support, workforce diversity, and governance), identifying both developments such as available data and tools and continuing obstacles including data collection difficulties, limited practical guidance, insufficient diversity in tech, and unclear regulatory expectations. The section identifies barriers to progress such as risks of penalising organisations for publicly addressing bias, insufficient transparency incentives, regulator unpreparedness, and industry nervousness about data protection law, and proposes anticipatory governance as an approach for organisations and regulators to anticipate and manage bias risks in advance rather than responding after deployment, with governance, transparency in the public sector, and regulatory clarity identified as key themes.
7 Enabling fair innovation
4 sections
Organisations need practical support to address algorithmic bias through better diversity, data on protected characteristics for monitoring outcomes, and technical methods for detecting and mitigating unfairness. Bias mitigation requires more than technical solutions and demands legal clarity on how to embed fairness into development and operational processes.
7.2 Workforce diversity
Technology sector workforce shows persistent gender imbalance (19% women over 10 years) and underrepresentation of Black people, with gaps in data on disability, age, socio-economic background and geographic diversity, though international representation is relatively strong; inclusive workplaces are necessary to retain talented staff from different backgrounds. Government supports diversity through the Tech Talent Charter (around 500 signatories aiming for 600 by end of 2020), a £1 million Digital Skills Innovation Fund (2018), and £10 million funding for underrepresented students in AI courses. The review recommends government continue supporting diversity programmes and advises organisations building algorithmic systems to prioritise workforce diversity across data science, operations, management and oversight roles, using data to identify and challenge recruitment and progression barriers.
7.3 Protected characteristic data and monitoring outcomes
The issue
Data protection concerns
The need for guidance
Alternative approaches
Access to baseline data
7.4 Detecting and mitigating bias
Multiple approaches exist to identify algorithmic bias, such as comparing training data with population datasets, analysing outcome differences between groups, and examining correlations between model variables and different groups. Appropriate methods vary by context, and monitoring is necessary whether decision-making uses algorithms or not.
Statistical definitions of fairness
Mitigating bias
Practical challenges
Legal and policy issues
7.5 Anticipatory Governance
Anticipatory governance requires organisations deploying algorithmic decision-making tools to proactively identify and intervene in potential issues before deployment through substantive impact assessment and consultation with affected parties, rather than treating these as tick-box procedures. Organisation leaders must ensure accountability for understanding tool capabilities and limits, considering fair treatment, deciding appropriate human involvement levels, gathering monitoring data, and understanding legal obligations, with Data Protection Impact Assessments, Equality Impact Assessments, and Human Rights Impact Assessments providing useful structures for planning. This approach is particularly important in the public sector where citizens often have no choice about service use and decisions can significantly affect individuals.
8 The regulatory environment
4 sections
Regulation can establish minimum standards and provide enforcement to address algorithmic bias, though existing laws and regulatory frameworks may not fully address novel challenges posed by artificial intelligence. The current system involves multiple regulators and industry bodies rather than a specialized regulator; however, coordination across equal treatment, personal data, and sector-specific rules is increasingly important for managing protected information and fairness beyond traditional protected characteristics.
8.2 Current landscape
The UK regulatory environment comprises multiple regulators with varying responsibilities and powers, some with explicit duties to address bias and discrimination while others must consider these issues when regulating their sectors. Algorithms operate within existing decision-making processes rather than replacing them entirely, so the existing regulatory framework must evolve rather than create a new algorithmic regulator; the Equality Act 2010, Human Rights Act 1998, Data Protection Act 2018, and GDPR provide the primary legal protections, with the EHRC and ICO serving as cross-cutting regulators, though sector-specific regulators also have important oversight roles. Sector regulators are essential for addressing algorithmic bias because decision-making is inherently sector-specific and algorithms typically form only part of broader regulatory processes; while some argue for a new cross-cutting algorithms regulator, enhanced coordination and alignment among existing regulators is a more appropriate response.
8.3 Legal background
Equality Act
Human Rights Act
Data Protection Law
Consumer protection and sector-specific legislation
Limitations of current legislation
Case law on the Equality Act
Regulations and guidance
8.4 The role of regulators
Algorithmic decision-making develops differently across sectors, with the EHRC and ICO holding explicit regulatory responsibilities alongside sector-specific regulators who must address algorithmic bias within their own mandates.
The Equality and Human Rights Commission
The Information Commissioner’s Office
Sector and specialist regulators
8.5 Regulatory tools
Beyond enforcement and guidance, regulators and industry should consider complementary tools such as regulatory sandboxes, impact assessments, audit mechanisms and certification schemes, and improved coordination to address algorithmic bias in organisations.
Regulatory sandboxes
Impact assessments
Audit and certification
Regulatory coordination and alignment
9 Transparency in the public sector
3 sections
Public bodies making decisions about individuals through algorithms should maintain or improve accountability to citizens. Although transparency mechanisms already exist, the increasing adoption of algorithmic decision-making requires ensuring such tools are transparent; suppliers typically involve external parties, but final responsibility for fair decisions rests with the public sector organisation. Government is recommended to mandate transparency obligations on public sector organisations using algorithms with significant influence on important decisions about individuals.
9.1 Identifying the issue
Why the public sector?
What is the problem?
What is the value of transparency?
9.2 Delivering public sector transparency
Public sector organisations using algorithms that significantly impact individual decisions should face mandatory requirements to publish information about their rationale for algorithmic use, algorithm type, its integration into decision processes, and safeguards for fair treatment. Government should conduct a project to define this obligation precisely and test implementation, grounded in principles of trustworthiness and intelligent disclosure.
What is in scope?
What information should be published?
How does this fit with existing transparency measures?
Managing the process of transparency
The relationship between transparency and explainability
9.3 Public sector procurement
Public sector bodies using external technology suppliers or contractors for algorithmic decision-making retain ultimate accountability to Parliament, the public and elected officials for fair and responsible decisions, regardless of supply chain arrangements. Transparency mechanisms must function across different supply chain models where reliance on external tools and libraries is common.
Supply chain models
Addressing the issue
10 Next steps and future challenges
The CDEI commits to supporting government, industry and regulators through guidance on applying equality law to algorithmic decision-making, developing technical standards for bias detection and mitigation, advancing transparency approaches, strengthening regulator coordination, fostering a skilled AI accountability ecosystem, and working with public sector organisations on practical governance structures. The CDEI's AI Monitoring function will continue observing algorithmic decision-making development and new forms of discrimination, referring issues to relevant regulators and to government where existing regulations do not apply; government should establish clear national leadership and coordination on algorithmic bias, and the CDEI expects to integrate bias and fairness as cross-cutting ethical issues into its wider future work.
Appendix A
Is this page useful?
Controls
What this policy requires, forbids, grants or penalises, one control at a time and in our own words. Each control links to the official text, which is the authority. This is not legal advice.
Drafted by the Index Drafted automatically from the official text, not yet reviewed by a person.Reviewed Checked by a person against the official text.
Themes covered 21 of 35
Choose a theme to see only its controls. About the themes
Risk and scope 4 of 4
Data and privacy 2 of 4
- Cross-border data transfer, not covered
- Copyright and training data, not covered
Transparency and records 3 of 4
- Labelling of synthetic content, not covered
Technical safeguards 3 of 4
- Accuracy, robustness and cybersecurity, not covered
Assurance and lifecycle 2 of 4
- Registration and notification, not covered
- Incident and breach reporting, not covered
Rights and protected groups 1 of 4
- Children and vulnerable groups, not covered
- Biometric identification and surveillance, not covered
- Whistleblowing, complaints and redress, not covered
Roles and organisation 5 of 6
- General purpose and foundation models, not covered
Enforcement and sectors 1 of 5
- Sector rules: health, not covered
- Sector rules: finance and insurance, not covered
- Sector rules: employment, not covered
- Sector rules: education, not covered
The full list of controls could not be loaded, so only the first 40 are shown.
Enabling fair innovation 5 controls
- ObligationApplies to: AnyoneDrafted by the Index
Organisations should identify and specify what fairness they seek to accomplish and establish how they will pursue it.
- ObligationApplies to: Public authorityDrafted by the Index
Sector bodies, regulators, standards organisations and government should create transparent rules on what constitutes fair algorithmic practice in different contexts.
- ObligationApplies to: AnyoneDrafted by the Index
All organisations must remain accountable for the fairness of their decisions whether those decisions are made through algorithms or structured human processes.
- ObligationApplies to: Public authorityDrafted by the Index
Government and industry should advance workforce diversity across roles involved in developing and deploying algorithmic decision-making systems and demonstrate measurable progress.
- ObligationApplies to: AnyoneDrafted by the Index
Organisations should deploy machine learning methods that measure and reduce algorithmic bias and discrimination.
The regulatory environment 2 controls
- ObligationApplies to: Public authorityDrafted by the Index
Public sector organisations should take reasonable steps upfront to evaluate whether their algorithmic systems could have a discriminatory impact and should detect algorithmic bias on an ongoing basis.
- ObligationApplies to: Public authorityDrafted by the Index
Regulators should clarify how existing laws, particularly the Equality Act 2010, apply to algorithmic decision-making to give organisations certainty.
Public sector transparency 5 controls
- ObligationApplies to: Public authorityDrafted by the Index
Government should impose a binding transparency requirement on all public sector organisations that use algorithms significantly influencing important decisions about individuals.
- ObligationApplies to: Public authorityDrafted by the Index
Government should scope the transparency obligation more precisely and pilot an implementation approach before full rollout.
- ObligationApplies to: Public authorityDrafted by the Index
Public sector organisations should proactively publish information on the decision to use an algorithm, the algorithm type, its role in the decision process, and measures taken to ensure fair treatment.
- ObligationApplies to: Public authorityDrafted by the Index
Procurement processes should be updated to maintain standards for fair algorithmic decision-making across the supply chain, including with technology suppliers and business process outsourcing providers.
- ObligationApplies to: Public authorityDrafted by the Index
Public bodies remain ultimately accountable for fair decision-making, even where suppliers are involved in developing or delivering algorithmic tools.
Next steps and future challenges 2 controls
- ObligationApplies to: Public authorityDrafted by the Index
Government should establish clarity on whether national coordination of algorithmic decision-making challenges sits in central government, a specific regulator, or another body such as CDEI.
- ObligationApplies to: AnyoneDrafted by the Index
Organisations building and deploying algorithmic decision-making tools should innovate responsibly and consider the implications for individuals and society.
2.6 1 control
- ObligationApplies to: DeployerDrafted by the Index
Organisations should design algorithmic decision-making systems to benefit people and the planet, grounded in respect for the rule of law, human rights and democratic values, with safeguards including human intervention where needed to ensure fairness and justice.
2.7 4 controls
- ObligationApplies to: DeployerDrafted by the Index
Organisations using algorithms for significant decisions should measure bias through data analysis to determine whether their decision-making processes contain historical patterns of unfairness.
- ObligationApplies to: DeployerDrafted by the Index
Organisations should address historical bias patterns identified in their data and design bias mitigation strategies when implementing algorithms to replace human decision-making systems.
- ObligationApplies to: DeployerDrafted by the Index
Organisations should assess the fairness of algorithms by measuring their impact on the overall decision process rather than evaluating algorithm outputs in isolation.
- ObligationApplies to: DeployerDrafted by the Index
Humans with supervisory responsibility for algorithms should understand how machine learning models work, their limitations and trade-offs, sufficiently to make informed judgements about their fairness and effectiveness.
3 2 controls
- ObligationApplies to: DeployerDrafted by the Index
Organisations should conduct Equality Impact Assessments to measure how their recruitment models perform across candidates with different protected characteristics, including analysis of those with multiple intersecting characteristics.
- ObligationApplies to: Public authorityDrafted by the Index
The Information Commissioner's Office should collaborate with industry to understand gaps in how current data protection guidance is applied in practice and should consider strengthening or promoting guidance as needed.
7.2 1 control
- ObligationApplies to: ProviderDrafted by the Index
Organisations should increase workforce diversity in technology teams to better identify and prevent biases being embedded or amplified in algorithmic systems.
The issue 2 controls
- ObligationApplies to: AnyoneDrafted by the Index
Organisations should measure whether their decision-making processes achieve fair outcomes by comparing results across different demographic groups.
- ObligationApplies to: AnyoneDrafted by the Index
Organisations should collect data on demographic characteristics to enable assessment of whether indirect discrimination is occurring in algorithmic systems.
The need for guidance 2 controls
- ObligationApplies to: Public authorityDrafted by the Index
Public sector organisations should collect and use protected characteristic data as essential practice for monitoring unintentional discrimination, consistent with the Public Sector Equality Duty.
- ObligationApplies to: Public authorityDrafted by the Index
Guidance should be updated and made consistent for public and private sector organisations on how to collect protected characteristic data for equality monitoring in algorithmic decision-making.
Alternative approaches 2 controls
- Applies to: AnyoneDrafted by the Index
Organisations may explore alternative models such as engaging a neutral third-party organisation or regulator to collect and test protected characteristic data on their behalf without providing the data to companies themselves.
- Applies to: AnyoneDrafted by the Index
When direct collection of protected characteristic data is unfeasible, organisations may infer protected characteristics from other available data solely for monitoring purposes, though they should carefully consider ethical, legal and trust implications.
Access to baseline data 1 control
- ObligationApplies to: Public authorityDrafted by the Index
Government and the ONS should make the Secure Research Service accessible to a broader range of organisations beyond academic researchers, including regulators and delivery organisations seeking to evaluate diversity outcomes.
Statistical definitions of fairness 1 control
- ObligationApplies to: ProviderDrafted by the Index
Organisations developing algorithms should select and apply a definition of fairness appropriate to their context, taking into account wider notions of fairness and recognising that no single mathematical definition applies universally.
Mitigating bias 1 control
- ObligationApplies to: ProviderDrafted by the Index
Organisations should implement protocols to detect and mitigate bias in statistical approaches by using pre-processing, in-processing or post-processing interventions at appropriate stages of algorithm development and deployment.
Practical challenges 1 control
- ObligationApplies to: DeployerDrafted by the Index
Organisations should learn what stakeholders and end users expect regarding fairness, the scope of impact and whether individuals retain agency before determining which fairness metrics to apply.
Legal and policy issues 2 controls
- ObligationApplies to: DeployerDrafted by the Index
Organisations must ensure that any bias mitigation techniques used align with UK equality law and data protection law.
- ObligationApplies to: DeployerDrafted by the Index
Organisations using fairness tools developed in the United States must recognise that those tools may not comply with UK or European equality law.
7.5 3 controls
- ObligationApplies to: DeployerDrafted by the Index
Organisations should implement governance structures proportionate to the risks posed by their use of automated decision-making that identify potential harms before systems are deployed.
- ObligationApplies to: DeployerDrafted by the Index
Leaders responsible for automated decision-making systems should consult with a diverse range of affected parties within and outside the organisation and conduct genuine impact assessments before deployment, not merely follow compliance procedures.
- ObligationApplies to: DeployerDrafted by the Index
Organisations should assess not only the technical details of how an algorithm operates but also whether it is suitable to use in the given circumstances and how it interacts with human decision-makers.
Human Rights Act 1 control
- ObligationApplies to: Public authorityDrafted by the Index
Public authorities must not discriminate on grounds including sex, race, colour, language, religion, political opinion, national or social origin, property or birth when securing rights and freedoms.
Data Protection Law 4 controls
- ObligationApplies to: ProviderDrafted by the Index
Organizations must process personal data lawfully, fairly and transparently in accordance with data protection principles.
- RightApplies to: OthersDrafted by the Index
Data subjects have rights to access information that organizations store about them and to learn how their data is used.
Change history
No changes recorded since we began monitoring this policy.