AEVOMIND INSIGHTS — PUBLIC SECTOR
AI in UK government and the public sector: the plan, the principles and how to implement it

The UK government wants AI used across public services and has published the guidance to do it: a 50-action plan, a playbook with ten principles, and a mandatory standard for publishing how algorithmic tools are used. Most public bodies are still early: the National Audit Office found 37% had deployed AI, while 70% were piloting or planning it.
The short answer
The UK government wants AI used across public services, and has published the guidance to do it: a national action plan, a playbook with ten principles, and a mandatory standard for publishing how algorithmic tools are used. Most public bodies are still early: the National Audit Office found just over a third had deployed AI, usually in one or two places, while most were piloting or planning.
This article summarises the plan, the principles and the transparency rules, and sets out a practical route for a public body — including colleges and schools — from a first pilot to a service people can trust.
Where government stands
When the National Audit Office reported in March 2024, 37% of government bodies responding to its survey had deployed AI and 70% were piloting or planning it. It identified skills, legal clarity, data quality and infrastructure among the barriers.
Since then the direction has been set from the top. The AI Opportunities Action Plan, published on 13 January 2025, set out 50 actions, all accepted by the government. In its one-year update the government said it had met its commitments on 38 of them.
Source: GOV.UK, AI Opportunities Action Plan: one year on (29 January 2026). Not yet met = 50 − 38.
Show the data
| Value | Share | |
|---|---|---|
| Commitments met | 38 actions | 76.0% |
| Not yet met | 12 actions | 24.0% |

The ten principles in the AI Playbook
The AI Playbook for the UK Government, published on 10 February 2025 for central government departments and public sector organisations, sets out ten principles:
- You know what AI is and what its limitations are.
- You use AI lawfully, ethically and responsibly.
- You know how to use AI securely.
- You have meaningful human control at the right stages.
- You understand how to manage the full AI life cycle.
- You use the right tool for the job.
- You are open and collaborative.
- You work with commercial colleagues from the start.
- You have the skills and expertise needed to implement and use AI solutions.
- You use these principles alongside your organisation’s policies and have the right assurance in place.

Transparency: the ATRS
The Algorithmic Transparency Recording Standard is mandatory for ministerial and non-ministerial departments, and for arm’s-length bodies that provide public or frontline services or routinely interact with the public. It applies to tools that have a significant influence on a decision-making process with public effect, or directly interact with the general public. Use became mandatory for most ministerial departments and HMRC on 6 February 2024, with further phases after.
Even where it is not mandatory, the record is a useful discipline: what the tool does, why, what data it uses, who is accountable and how people can challenge it.
Automated decisions about people
Under the UK GDPR as amended by the Data (Use and Access) Act 2025, a significant decision about a person based solely on automated processing must come with safeguards: information about the decision, the chance to make representations, human intervention, and the right to contest it. Special category data remains more tightly restricted. For most public services, the safer design is the one the Playbook describes: meaningful human control at the right stage. We cover the detail in the UK rules on AI.

The arithmetic of a public enquiry service
Take an illustrative public body handling 3,000 routine enquiries a month — opening times, how to apply, where a case has got to — at about six minutes each. That is 300 officer hours a month. Suppose an AI assistant answers half of them fully from approved information, with an officer checking each answer, and prepares the rest so an officer needs half the time.
Illustrative inputs, not data from any public body: 6 minutes per enquiry by hand; half answered fully by the assistant from approved information, with officers checking each answer at about 30 seconds; the rest prepared by the assistant so an officer needs 3 minutes. Totals: 300 hours a month before, 87.5 after.
Show the data
| Before | With an AI assistant | |
|---|---|---|
| Routine enquiries answered by AI | 150.0 hrs | 12.5 hrs |
| Enquiries that need an officer | 150.0 hrs | 75.0 hrs |
That is 300 hours a month before against 87.5 after — about 212 hours returned to casework, roughly 1.3 full-time officers at 162.5 hours a month. The inputs are illustrative; the shape is typical. The value is not fewer staff but faster answers for the public and more time for the cases that need judgement.
A practical route for a public body
- Start with a service pain point, such as a backlog of routine enquiries, not with a technology.
- Measure it: volumes, handling time, waiting time, error and complaint rates.
- Do the data protection impact assessment early; for AI it is required in most cases.
- Design human control in: who checks what, and how the public can reach a person.
- Bring commercial colleagues in from the start, as the Playbook says.
- Publish an ATRS record if the tool is in scope, and keep an audit trail of what it did.
- Pilot, measure against the baseline, and only then scale.
The same route applies to colleges and schools; see software for colleges and software for schools. For how an AI assistant holds a defined job with escalation rules, see what AI employees are.

Frequently asked questions
What is the AI Playbook for the UK Government?
Published on 10 February 2025, it is government guidance on using AI safely and effectively in central government departments and public sector organisations. It sets out ten principles, from knowing what AI is and its limits, through using it lawfully, ethically and securely with meaningful human control, to working with commercial colleagues from the start and having the right assurance in place.
What is the Algorithmic Transparency Recording Standard?
The ATRS is a standard way for public sector bodies to publish information about the algorithmic tools they use and why. It is mandatory for ministerial and non-ministerial departments, and for arm's-length bodies that provide public or frontline services or routinely interact with the public, for tools that significantly influence decisions with public effect or interact directly with the public.
How many government bodies use AI?
When the National Audit Office reported in March 2024, 37% of the government bodies responding to its survey had deployed AI, typically in one or two use cases, and 70% were piloting or planning AI. The report found that skills, legal clarity, data quality and infrastructure were among the barriers to wider use.
What is the AI Opportunities Action Plan?
Published on 13 January 2025, it set out 50 actions to grow the UK's AI capability and use AI across the economy and public services; the government agreed to take forward all of them. In its one-year update on 29 January 2026, the government said it had met its commitments on 38 of the 50 actions.
Can a public body make automated decisions about people?
It can, within the law. Under the UK GDPR as amended by the Data (Use and Access) Act 2025, a significant decision based solely on automated processing must come with safeguards: telling the person about the decision, letting them make representations, letting them obtain human intervention and letting them contest it. Special category data remains more tightly restricted.