A Data Protection Impact Assessment (DPIA) is a structured process for identifying and reducing privacy risks before an organisation starts high-risk processing of personal data. Under the UK GDPR, it is required where processing is likely to result in a high risk to individuals, such as large-scale monitoring, profiling, use of special category data, biometric identification or deployment of new technology. Harper James is a UK full-service commercial law firm advising businesses on data protection and the laws of England and Wales.
A DPIA should begin during project design, not after a system or process has been implemented. Start by describing the processing, its purpose, the categories of personal data and individuals involved, data flows, retention periods, access controls and any suppliers or international transfers. The assessment should explain why the processing is necessary and proportionate, and record the legal basis relied on.
Identify the risks to individuals, including loss of confidentiality, inaccurate decisions, discrimination, excessive retention, lack of transparency and inability to exercise data rights. Then set out specific controls, such as data minimisation, pseudonymisation, encryption, access restrictions, human review, retention limits, staff training and supplier obligations. Consult the Data Protection Officer where one is appointed, and involve security, procurement, product and operational teams so that the assessment reflects the real workflow. The ICO’s DPIA guidance and templates provide a useful reference for documenting the analysis.
The final document should state the residual risk after controls are applied, assign owners and include review dates or trigger events. Revisit the DPIA when the purpose, technology, data sources, suppliers or risk profile changes. If high residual risk remains and cannot be reduced, the organisation must consult the Information Commissioner’s Office before processing begins. A DPIA is therefore a working governance record, not a form filed once and forgotten.
Projects involving generative artificial intelligence, automated decision-making, employee monitoring, connected devices, facial recognition and large datasets require careful scrutiny of purpose, accuracy, transparency and human oversight. Organisations should also test whether their privacy notice, contracts, security measures, retention policy and procedures for subject access requests match the processing described in the DPIA. A clear assessment supports better decisions by showing which risks are accepted, which controls are mandatory and who is accountable for delivery.