Before using AI tools with personal data, confirm your lawful basis, complete a risk assessment where needed, track where data goes, and keep people involved in key decisions. This guide outlines the regulator's rules, but it is not legal advice.
The Data (Use and Access) Act 2025 updated UK data protection rules. For significant automated decisions, the ICO confirms you can now use the full range of lawful bases, including legitimate interests, provided you put appropriate safeguards in place. These changes do not apply to special category data, which has stronger protection.
The ICO's main guidance on AI and data protection dates from 2023 and is currently under review because of the Act, so do check for updates. This article provides general guidance, not legal advice.
A lawful basis for each use
The ICO guidance says that processing personal data to train an AI system or make predictions requires an appropriate lawful basis. In practice, list each way an AI tool uses personal data and note the basis for each.
Assess risk first
The ICO notes that in the vast majority of cases, using AI involves processing likely to result in high risk to individuals' rights and freedoms. This triggers a legal requirement to carry out a data protection impact assessment, assessed case by case.
If you cannot assess whether a third-party AI tool would be compliant, the ICO says you should choose a different one as good practice.
Know where your data goes
Many AI tools are run by companies outside the UK. The ICO's international transfer rules state that restricted transfers need a valid mechanism, such as UK adequacy regulations, appropriate safeguards or an exception, alongside a transfer risk assessment. Whether a tool involves a restricted transfer depends on how the supplier is set up, so ask them.
Keep people in significant decisions
Under the Data (Use and Access) Act 2025, significant decisions made solely by automation need safeguards. People must be able to contest them and get human intervention. In 2026, an ICO review of over 30 employers found many likely used automated recruitment without meaningful human involvement.
What to check first
- 01Will the tool handle any personal data?
- 02What is your lawful basis for each use?
- 03Have you assessed the risk, with a DPIA if needed?
- 04Where is the data stored and processed?
- 05Does the supplier use your data to train its models?
- 06Can a person review and change significant decisions?
- 07Have you checked the ICO's latest guidance?
Our AI Readiness Audit includes a note on how your data is handled, written to stand up to an auditor's questions. We are not lawyers. For legal advice, you should speak to a qualified adviser.
Sources
Get our newest articles
We only email you when a new piece goes live.



