AI to cut paperwork to free up doctors' time for patients but let's dig deeper as to who has your data - American companies?

17th August 2025

Patients and frontline staff could see huge benefits from new AI helping people out of hospital quicker and slashing bureaucracy.

Patients and frontline staff could see huge benefits from new AI helping people out of hospital quicker and slashing bureaucracy.

Tool is one of the Prime Minister's AI Exemplars, including real-world projects using AI to make people's lives easier and modernise services across health, justice, tax and planning.

Group of leading projects will receive support to expand the use of their technology more quickly, helping to drive efficiencies and boost growth through Plan for Change.

Patients could get home to family and off busy wards more quickly, thanks to game-changing AI that could help write the documents that are needed to discharge people from hospital.

The cutting-edge technology will help cut waiting lists, by giving frontline staff the precious gift of time and making care more efficient so that loved ones return to the comfort of their homes quickly. Currently being developed at Chelsea and Westminster NHS Trust, it is one of many projects to receive backing from the Prime Minister as part of the AI Exemplars programme.

The AI-assisted tool could deliver the support that NHS staff have been crying out for - helping doctors to draft discharge documents faster by extracting key details from medical records, such as diagnoses and test results, using a large language model. After a full review from a medical expert responsible for the patient, these documents are then used to discharge a patient from a ward and refer them to other care services that may be needed.

It would radically improve an outdated system that can leave patients on wards unnecessarily for hours, waiting for time-pressed doctors providing urgent care to sit down and fill in forms before they can go home. In some cases, the current system for writing discharge summaries can also inaccurately record basic patient details - like what treatment they've had, or changes to medication - and put them in harms way.

Another project announced today, ‘Justice Transcribe', will be transformational for Probation Officers - by helping to transcribe and take notes in their meetings with offenders after they leave prison. The technology, which was found to halve the time officers spent organising notes between meetings and in their personal time, is set to be given to all 12,000 probation officers later this year.

Projects being announced today as part of the Prime Minister's AI Exemplars programme are prime examples of how the government wants to use AI across the public sector to make people's lives easier and help deliver the Plan for Change. Over the coming months, these exemplars will be developed and trialled, with those showing the most promise potentially rolled out more widely. It follows the Prime Minister's approach that people should not spend their time on tasks that AI can do quicker and better.

Speaking on a visit to Chelsea and Westminster Hospital, Technology Secretary Peter Kyle said:

This is exactly the kind of change we need: AI being used to give doctors, probation officers and other key workers more time to focus on delivering better outcomes and speeding up vital services.

This government inherited a public sector decimated by years of under-investment and is crying out for reform. These AI Exemplars show the best ways in which we're using tech to build a smarter, more efficient state.

When we get this right across government, we're talking about unlocking £45 billion in productivity gains - delivering our Plan for Change and investing in growth not bureaucracy.

Health and Social Care Secretary Wes Streeting said:

This potentially transformational discharge tool is a prime example of how we're shifting from analogue to digital as part of our 10 Year Health Plan.

We're using cutting-edge technology to build an NHS fit for the future and tackle the hospital backlogs that have left too many people waiting too long.

Doctors will spend less time on paperwork and more time with patients, getting people home to their families faster and freeing up beds for those who need them most.

The NHS Federated Data Platform, a system designed to connect IT across health and care services, is hosting the AI-assisted discharge summaries tool. This means that it can handover information to different care services in an efficient and secure way, while also making it easier to use the technology across the country if tests are successful.

Planning

The AI Exemplars programme will also include the ‘Extract' tool, which will standardise data faster by converting decades-old, handwritten planning documents and maps into data in minutes. It will power new types of planning software to slash the 250,000 estimated hours spent by planning officers each year manually checking these documents.

Schools

Other technology backed by the programme, the ‘AI Content Store', will also help make more accurate AI tools to support teachers to mark work and plan lessons - ensuring they are able to spend more time helping children in the classroom with face-to-face teaching, supporting the government's mission to break down barriers to opportunity.

Justice

A further tool in the programme is ‘Justice Transcribe'. Early feedback from probation officers has shown that the technology allows them to focus on the personal, and often emotive meetings with offenders, instead of having to interrupt to take notes and clarify details. The technology is based on ‘Minute', part of the Humphrey package of AI tools built by government to help make the civil service more efficient.

Civil service

The suite of AI tools known as ‘Humphrey', that helps make the civil service more efficient, is also included in the package. It comes as ‘Consult', a tool in the package, analyses the thousands of responses any government consultation might receive in hours, before presenting policy makers and experts with interactive dashboards to explore what the public are saying directly.

It has been the first AI tool to undergo testing against a new ‘social readiness' standard, where the tech was shared with members of the public to get their views on the value it adds, the strength of safeguards in place and the risks associated with using the technology. Members of the public noted that Consult is well targeted to replace an "old school process" that is very "archaic" and ripe for improvement with AI.

The independent report, completed after deliberative focus groups by the Centre for Collective Intelligence at Nesta, a charity focused on innovation for the public good, found that 82% of people felt positive or neutral about the use of the technology across government.

With more to be announced in the coming months, AI Exemplars include:

Justice Transcribe, Ministry of Justice.
‘Humphrey', Department for Science, Innovation and Technology.
Education Content store, Department for Education.
AI Tax Compliance, HMRC.
‘Extract' and the Digital Planning Programme, Department for Science, Innovation and Technology and Ministry of Housing, Communities and Local Government.
‘Minute' for Local Government, Department for Science, Innovation and Technology.
GOV.UK Chat, Department for Science, Innovation and Technology.
AI for diagnostics, NHS.

We are all being outsourced now
Here's a short, sourced vendor list for each department (NHS, Home Office, DWP).

NHS (examples of vendors used recently)

Palantir — won the NHS England Federated Data Platform / large data-platform work (controversial, widely reported).

Microsoft / Azure — cloud/AI platform supplier across Whitehall and cited as a public-sector cloud partner for NHS projects.

Amazon Web Services (AWS) — named as a cloud partner in NHS data platform work alongside other suppliers.
Palantir Blog

Faculty (UK) — London AI firm contracted into Covid/NHS data work and other NHS analytics projects.

DeepMind / Google Health (Alphabet/UK-founded DeepMind) — historic NHS partnerships (Streams app / data agreements) and subsequent scrutiny.

Home Office (examples / notable suppliers & activity)

Palantir — awarded/used for border and customs data tooling (post-Brexit border work), and referenced in multiple Home Office/cabinet-office contracts and reporting.

Major cloud/tech vendors & consultancies (Microsoft, AWS, Deloitte, KPMG etc.) — Home Office uses major cloud providers and large consultancies on digital and AI projects; recent procurement activity shows Home Office seeking AI-as-a-service / AI infrastructure partners via the Crown Commercial DPS.

In-house/algorithmic systems — Home Office has used internal/contracted systems (e.g., case-prioritisation/IPIC) that rely on automated scoring and analytics; these systems have drawn scrutiny and calls for greater transparency.

DWP (Department for Work & Pensions)

Various AI suppliers via CCS AI DPS (RM6200) — DWP has initiated multiple AI procurements (e.g., Nexus AI / gen-AI lighthouse projects) and is running tenders through the Crown Commercial Service's AI DPS, meaning many vendors (commercial AI firms, cloud providers and consultancies) are eligible/being considered.

Palantir (lobbying / engagement) — Palantir has actively marketed fraud-detection tech to DWP and lobbied ministers (reported).

Microsoft (via resellers / large contracts) — DWP has large Microsoft software arrangements (e.g., reseller deals) and therefore uses Microsoft cloud/AI tooling indirectly in parts of its estate.

In-use AI systems for fraud detection / casework — DWP uses machine-learning systems for fraud and benefits administration; coverage has flagged fairness/bias issues in some deployed models.

Notes
Palantir is an American company.
Founded in 2003 in Palo Alto, incorporated in Delaware, and now headquartered in Denver, Colorado. It's a publicly traded U.S. company (NYSE: PLTR) and operates internationally, including offices and contracts in the UK.

DeepMind is now owned by Alphabet Inc
DeepMind was founded in London in 2010 and bought by Google in 2014.
TechCrunch
Wikipedia

Today DeepMind operates as a subsidiary within Alphabet (Google's parent company), though it remains headquartered in London and keeps research sites outside the US.
(Related note) DeepMind’s health arm was folded into Google Health in 2018 — a move often cited when people discuss how DeepMind’s work got more closely integrated with Google/Alphabet.

Faculty is a privately-held UK company — not American-owned but its a minnow compared to the American companies.
aculty was founded in London in 2014 by Marc Warner, Angie Ma and Andrew (Andy) Brookes and those founders remain senior directors.
Wikipedia
faculty.ai

It is privately owned: in May 2021 Faculty raised a £30m growth round led by the Apax Digital Fund (Apax Partners), so Apax is a major investor but not an American parent company.
Other investors named in reporting and filings include Guardian Media Group Ventures, LocalGlobe and individual investors — so ownership is a mix of founders + institutional/private investors.
Faculty is tiny compared with Meta or Alphabet/Google.
Faculty is a small, privately-held UK AI consultancy (hundreds of employees, tens of millions in estimated revenue) versus Meta and Alphabet which have tens-hundreds of thousands of employees, annual revenues in the hundreds of billions, and market caps in the trillions.
aculty is a specialist consultancy / systems integrator with a few hundred staff and modest revenues compared with global tech giants.
Dealroom.co
Growjo

Meta and Alphabet are massive platform companies: they operate global ad/cloud/consumer businesses, employ 10⁵-10⁶ people scale across divisions, and generate revenues and market valuations thousands of times larger than a typical private AI consultancy.

What are the dangers of outsourcing UK health data to foreign companies like alphabet and Meta
Outsourcing UK health data to large foreign tech firms creates real benefits (scale, tooling, expertise) but also several concrete risks — legal, privacy, security, commercial and political. Below list of the the main dangers, why they matter, and practical mitigations the NHS / government typically use.

Foreign legal access / cross-border disclosure
Data held by US companies can be subject to US legal powers (eg. the CLOUD Act and related data-access agreements), which can require disclosure to foreign law-enforcement even if the data was collected in the UK. That creates a risk that sensitive health data could be accessed under a foreign legal process.
Department of Justice

Historical failures of lawful, transparent handling
Real cases show how agreements went wrong — e.g. the ICO found problems in the Royal Free / DeepMind arrangement and required remedial action; those episodes underpin concerns about insufficient notice, scope creep and inadequate safeguards. Past incidents erode public trust.
Information Commissioner's Office
The Guardian

Security and operational risk (breaches, misconfiguration, supply-chain)
Large cloud or platform deployments reduce some risks but introduce others (misconfigurations, privileged-access compromise, vendor supply-chain attacks). UK guidance emphasises cloud security principles and careful architecture to manage these risks.
NCSC

Loss of control & vendor lock-in
If critical datasets, pipelines or models live on a vendor’s platform, the NHS can become dependent on that vendor’s pricing, APIs and roadmaps — making it hard to move data or switch providers without cost, delay or fragmentation. Parliamentary scrutiny has explicitly highlighted concerns about dependence on particular suppliers for national systems.
UK Parliament Committees

Commercialisation and secondary uses
When external firms host or process health data there’s risk that the data (or insights from it) could be reused or monetised in ways not originally intended — even if contracts forbid it, weak contract terms or poor oversight can allow scope creep. This is also a public-trust problem.
Information Commissioner's Office
The Guardian

AI bias, opacity and damaged care outcomes
Outsourced models and tooling may be opaque (black-box), trained on data that don’t reflect UK populations, or produce biased outputs. That can cause unfair treatment, wrong clinical decisions or unequal outcomes if not carefully validated and governed. (See below for mitigations and governance.)
NHS England Digital
NCSC

Practical mitigations / what the UK does or should require
Data minimisation & pseudonymisation — only share the minimum data necessary; use robust pseudonymisation before external processing and split keys/identifiers where possible. (ICO / NHS guidance covers these.)

Strong contracts + DPIAs + legal safeguards — data processing agreements that ban secondary uses, demand audits, logging, breach notifications and clear exit/repatriation clauses; complete Data Protection Impact Assessments (DPIAs).

Technical controls & cloud security principles — apply NCSC cloud security principles: least privilege, strong encryption (at rest and in transit), provenance and robust key management, network isolation and HSCN connectivity rules for health services.

Transparency & public engagement — publish what’s shared and why, offer opt-outs where appropriate, and be transparent about third-party relationships to maintain trust. Parliamentary scrutiny and regulators expect clear transparency.

Preferential/sovereign hosting where needed — for the most sensitive datasets, governments sometimes require UK/EU hosting, or use "sovereign cloud" contracts that limit cross-border access and include contractual promises about government requests. NHS guidance on off-shoring/cloud explains safeguards to allow safe use of public cloud where appropriate.

Independent validation of models — clinical validation, fairness testing, explainability requirements and ongoing monitoring before deployment in patient care.

In plain terms

Working with big foreign tech firms can accelerate capabilities, but it raises non-trivial legal, security and governance issues. Those risks can be managed — but only if organisations insist on strong technical controls, legally enforceable contracts, independent oversight, and transparency to patients and the public. UK regulators and NHS guidance now expect these controls to be in place for off-shore/public-cloud arrangements.

A House of Commons committee recently took evidence on this topic -
House of Commons Science Innovation and Technology Committee 8 July 2025