Why this course starts with a map
Organisations now use computer systems, including artificial intelligence (AI), to sort, score and rank people. You will often not be told. Before you can use your rights, you need to spot where these systems are used. In this lesson you will:
- Map where AI may make or shape decisions about you.
- Learn to tell a fully automated decision from one where AI only helps a person decide.
- Understand what this course can and cannot do for you.
- Learn how to check a legal point and record where and when you read it.
Important: This course gives general information about UK law. It is not legal advice. The law applies differently to different facts, and it changes. If you have a real dispute, see the section "When to get professional help" below.
Task 1: Map where AI touches your life
Think of three roles you play: worker, customer and citizen. In each one, an organisation may use automated tools on you.
At work, or looking for work
- CV screening software that ranks or rejects applications.
- Online aptitude tests or video interviews scored by an algorithm.
- Monitoring tools that track productivity, and systems that flag workers for performance reviews or allocate shifts.
As a customer
- Credit checks and online loan or credit card applications that are approved or refused in seconds.
- Insurance quotes priced using data about you, your area or your behaviour.
- Fraud checks that block a payment or freeze an account.
- Customer service chatbots that answer questions or refuse requests.
As a citizen
- Benefit claims checked by tools that flag cases for extra investigation.
- Health services that use online triage or tools that help read scans.
- Policing and immigration systems that help assess risk (covered in outline later in the course).
The UK government asks public sector bodies to publish records of the algorithmic tools they use under the Algorithmic Transparency Recording Standard. These records are on GOV.UK.
DO: Write down five situations from the last two years where a decision about you may have involved an automated system. Next to each, note the organisation and the outcome.
Task 2: Decide whether AI made the decision or only helped
This difference matters because UK data protection law has specific rules for decisions made without meaningful human involvement. The Information Commissioner's Office (ICO), the UK data protection regulator, calls these "solely automated" decisions. Its examples include an online credit application that is refused automatically and recruitment tests that use pre-programmed algorithms.
There are two basic types:
- Fully automated: the system decides and the result goes straight to you. Example: your online loan application is refused within ten seconds and no person ever looks at it.
- Decision support: the system produces a score or recommendation and a person makes the final decision. Example: software ranks 300 CVs and a recruiter reads the top 40 before choosing who to interview.
The line can blur. According to the ICO, human involvement only counts if it is real: the person must be able to change the outcome and must actually consider it. A manager who approves every system recommendation without looking at it is not adding meaningful involvement.
DO: For each of your five situations, ask:
- Did a person review my case before the decision was made?
- Could that person have changed the outcome?
- Did they actually weigh up the facts, or just accept the system's output?
Mark each one as "fully automated", "AI-supported" or "don't know". You will learn how to find out in later lessons.
Task 3: Know the limits of general information
Think of this course as a road map. It shows you the routes. It does not drive the car for you. It cannot:
- apply the law to the full facts of your case;
- tell you whether you will win a claim;
- keep track of deadlines that apply to you.
When to get professional help
Get advice from a solicitor, Citizens Advice, a trade union or another qualified adviser if:
- you are thinking of going to a court or employment tribunal (time limits are strict and can be short);
- you have lost a job, a large sum of money, housing or a benefit;
- the other side has instructed lawyers;
- you believe you have been discriminated against.
Task 4: Check a legal point and record it
Law changes. For example, the Data (Use and Access) Act 2025 received Royal Assent on 19 June 2025. It changed the rules on automated decision-making, but those changes did not take effect that day. They came into force on 5 February 2026 under commencement regulations. An article written in 2024 would therefore be out of date.
Use this procedure every time:
- Find the primary source. For the law itself, use legislation.gov.uk. For guidance on data rights, use the ICO website (ico.org.uk). For government policy and services, use GOV.UK.
- Check it is current. On legislation.gov.uk, read the "Changes to legislation" notice. It tells you whether the text is up to date with changes in force on a given date, or whether there are changes not yet shown. On ICO pages, look for notices that guidance is "under review".
- Check commencement. An Act being passed does not mean every section is in force. Look for commencement regulations.
- Record it. Write down the title of the page, the URL and the date you read it.
Example record:
Point: Rules on automated decisions changed by the Data (Use and Access) Act 2025
Source: legislation.gov.uk, SI 2026/82 (commencement regulations)
URL: https://www.legislation.gov.uk/uksi/2026/82/made
Date read: 22 October 2026
Checklist: before you move on
- I have listed five situations where AI may have shaped a decision about me.
- I have marked each as fully automated, AI-supported or unknown.
- I know this course is general information, not legal advice.
- I know when to get professional help.
- I have bookmarked legislation.gov.uk, ico.org.uk and GOV.UK.
- I have a notes page that records source, URL and date read.
Key takeaways
- AI can affect you as a worker, a customer and a citizen: in hiring, credit, insurance, fraud checks, benefits and health.
- A fully automated decision has no meaningful human involvement. With decision support, a person really reviews the case and can change the outcome.
- A person who rubber-stamps a system's output does not count as meaningful human involvement.
- This course is general information, not legal advice. Get professional help for disputes, deadlines and serious losses.
- Always check the primary source and its commencement date, and write down the URL and the date you read it.