Last reviewed: Next review due:
What the technology beat covers
The technology beat in the UK spans two distinct but overlapping worlds. The first is the domestic tech ecosystem: the startups, scale-ups, venture capital flows, university spin-outs, and government funding programmes tracked by publications like UKTN, TechCrunch UK, and the former Tech Nation reports. The second is the global tech giants — Google, Meta, Apple, Amazon, Microsoft — whose decisions about products, pricing, and content moderation have direct effects on millions of UK citizens and whose UK regulatory exposure is growing rapidly.
AI has become its own sub-beat since 2022. This requires journalists who can distinguish genuine capability advances from marketing, explain large language model limitations to general audiences, probe algorithmic systems for bias or harm, and hold regulators to account for the pace of their response. The AI Safety Institute (AISI), rebranded as the AI Security Institute, publishes frontier model evaluations that are primary source material for any AI reporter.
A third pillar is data and surveillance: ICO enforcement actions, data breach notifications, biometric surveillance in public spaces, and the intersection of intelligence capabilities with consumer technology. This overlaps significantly with the civil liberties and security beats.
Why this beat matters
- 1The Online Safety Act 2023 represents the most significant intervention in internet governance since the web's creation. Ofcom's implementation decisions will shape what content 40+ million UK internet users can access.
- 2AI systems are being embedded into public-sector decision-making — benefits assessment, policing prediction tools, NHS triage — often without adequate transparency or independent audit.
- 3The CMA's Digital Markets, Competition and Consumers Act 2024 powers over Strategic Market Status firms could fundamentally reshape how Big Tech operates in the UK.
- 4Deepfake technology is being weaponised against politicians, journalists, and private individuals — intimate image abuse via AI is now a criminal offence under the Criminal Justice Act 2024.
- 5Tech investment decisions shape regional economic inequality: which cities get data centres, which do not; which communities benefit from AI productivity gains.
- 6Misinformation amplification by algorithmic recommendation systems intersects with every other journalism beat.
Core legal and ethical risks
Defamation
Startup founders and tech executives are increasingly litigious. Allegations of fraud, data misuse, or misconduct carry serious harm potential. See our defamation checklist.
Computer Misuse Act 1990
Accessing exposed data via an unprotected URL may constitute unauthorised access. Always take legal advice before accessing data discovered via security research.
AI-generated content accuracy
Using AI tools to draft copy creates real risk of hallucinated quotes, statistics, and case citations. IPSO Clause 1 accuracy obligations apply to everything you publish.
Source protection in tech reporting
Tech company employees who whistleblow face aggressive NDAs and litigation threats. Understand your obligations under RIPA/IPA before digital source communication.
Useful UK public datasets
FOI ideas for the technology beat
See our full FOI story ideas guide for drafting and submission tips.
- Home Office: procurement contracts for facial recognition technology, including supplier names, costs, and evaluation methodology.
- DWP: use of algorithmic tools in Universal Credit fraud detection — accuracy rates, false-positive rates, and appeal outcomes.
- NHS trusts: any AI diagnostic tools in clinical use — procurement basis, clinical trial data, and post-deployment audit results.
- Department for Education: EdTech contracts, data-sharing agreements with education software providers.
- Met Police: number of live facial recognition deployments 2022–2025, locations, and matches-to-arrests ratios.
- DSIT: all communications with Big Tech lobbyists regarding the Online Safety Act implementation.
Key UK source organisations
Ofcom
Online Safety Act regulator
ICO (Information Commissioner's Office)
Data protection and privacy enforcement
CMA
Digital markets competition
AI Safety Institute / DSIT
Frontier AI evaluation and policy
Alan Turing Institute
AI research and policy analysis
Open Rights Group
Digital rights campaigning
Big Brother Watch
Surveillance and civil liberties
techUK
Tech industry trade body
BCS — The Chartered Institute for IT
Professional IT standards
Nesta
Innovation policy think tank
Interview question bank
- Q1.What independent evaluation of this system's accuracy has been conducted, and by whom?
- Q2.Has a bias or equality impact assessment been completed before deployment? Can I see it?
- Q3.Which human is ultimately accountable if this AI system makes a decision that harms someone?
- Q4.What recourse do affected individuals have to challenge an automated decision?
- Q5.What data was used to train this model, and was consent obtained from data subjects?
- Q6.Has this system ever produced a false positive that led to real-world harm? What happened?
- Q7.What safeguards prevent the system being used beyond its stated purpose?
- Q8.How does your company's lobbying activity align with your stated AI safety principles?
Jargon glossary
- Large language model (LLM)
- A type of AI trained on vast text datasets to generate plausible text. Does not "understand" — predicts probable next tokens. Prone to hallucination.
- Hallucination
- When an AI system generates factually incorrect output presented with apparent confidence. A significant risk for AI-assisted journalism.
- Strategic Market Status (SMS)
- Designation under DMCCA 2024 by the CMA for firms with entrenched market power in digital activities, enabling conduct requirements.
- Online Safety Act (OSA)
- UK legislation imposing duties of care on platforms regarding illegal and harmful content, with Ofcom as enforcer.
- Algorithmic accountability
- The principle that automated decision-making systems should be explainable, auditable, and subject to meaningful human oversight.
- Deepfake
- AI-synthesised media — video, audio, images — depicting someone doing or saying something they did not. Creating intimate deepfakes is now a criminal offence.
- ICO
- Information Commissioner's Office — UK data protection regulator with powers to fine organisations for GDPR/UK GDPR breaches up to £17.5m or 4% of global turnover.
- Computer Misuse Act 1990
- UK law criminalising unauthorised access to computer systems and data. Relevant to security journalists who access exposed data.
- AI Safety Institute (AISI)
- UK government body responsible for evaluating risks from frontier AI models, now operating under the AI Security Institute branding.
- Digital Markets, Competition and Consumers Act (DMCCA) 2024
- UK legislation expanding CMA powers over digital markets, including SMS designation and pro-competition interventions.
Story ideas
- Map all algorithmic tools currently in use across DWP, HMRC, and Home Office — which have been independently audited?
- Track CMA Strategic Market Status designations: which companies have been designated, what conduct requirements are proposed, and are they complying?
- Investigate the rollout of live facial recognition by UK police forces — compare accuracy claims with operational data obtained via FOI.
- Profile the UK AI safety research ecosystem: who funds it, what conflicts of interest exist between commercial AI labs and safety researchers?
- Examine NHS AI diagnostic tool procurement: are any tools in clinical use that have not received MHRA approval as medical devices?
- Investigate the gap between tech company Online Safety Act compliance claims and independent researcher findings about harmful content moderation.
- Follow the money in UK AI investment: which venture funds are backing frontier AI startups and what are their ties to defence and intelligence?
- Assess deepfake detection capability in UK newsrooms — are publications equipped to verify AI-generated content before publication?
Pitch angles
- The human cost angle: find individuals whose lives have been materially affected by an algorithmic decision — benefit denial, wrongful fraud flags, biometric misidentification.
- The accountability gap angle: identify an AI system in public-sector use for which no one can explain the decision logic or accept accountability for errors.
- The regulatory lag angle: compare the pace of AI deployment in a specific sector with the pace of regulatory guidance and enforcement action.
- The hype vs reality angle: take a specific AI capability claim and test it empirically or via independent expert assessment.
Recommended tools
See the full verification tools and FOI tools sections in our tools directory.
- Companies House — verify AI startup registration and director details
- WayBack Machine — capture and preserve tech company claims before deletion
- FOI Directory (WhatDoTheyKnow) — search prior tech-related FOI requests
- CLIP/Google Lens — reverse-search AI-generated images
- InVID / WeVerify — video verification for deepfake detection
Related UK organisations
Related guides
Defamation Risk Checklist
Pre-publication defamation risk checklist for UK journalists.
Journalism Ethics
IPSO Code, NUJ Code, and ethical obligations for digital journalism.
FOI Story Ideas
How to generate investigations using Freedom of Information requests.
Media Law Hub
Full coverage of UK media law including privacy, contempt, and data protection.
Primary sources
Frequently asked questions
How do I verify claims made by AI companies about model capabilities?
Request third-party benchmark results, check peer-reviewed papers, and consult independent researchers at the Alan Turing Institute or university AI labs. Many vendor capability claims are marketing language. Look for reproducible evaluation methodology and ask whether claims have been independently replicated. The AI Safety Institute publishes frontier-model evaluations that provide a baseline.
What is the Online Safety Act and what does it require of platforms?
The Online Safety Act 2023 places duties of care on user-to-user platforms and search services to protect users from illegal content and, for larger platforms, legal but harmful content. Ofcom is the regulator and has sweeping powers including fines of up to £18m or 10% of global turnover. Platforms must assess risks, publish transparency reports, and give users safety tools. The Act also contains specific provisions on CSAM, fraud, and intimate image abuse.
Can I use AI-generated content in my journalism?
IPSO has issued guidance indicating that use of AI-generated content does not automatically breach the Editors' Code, but the accuracy obligations of Clause 1 still apply in full — the journalist and publisher are responsible for everything published. Disclosure to readers of AI involvement is increasingly expected as an ethical standard. Be particularly cautious about AI-generated quotes, statistics, or case law which may be hallucinated.
What legal risks apply specifically to the tech beat?
Defamation risk is high when reporting on startup founders accused of fraud or misconduct. Privacy risk arises from reporting on data breaches (misuse of private information). The Computer Misuse Act 1990 creates risk if a journalist accesses a system beyond authorisation — even clicking an exposed URL can be legally ambiguous. Contempt risk arises from active court proceedings involving tech companies. See our defamation and privacy law guides for pre-publication checklists.
What is the CMA's role in tech journalism?
The Competition and Markets Authority has designated powers under the Digital Markets, Competition and Consumers Act 2024 to designate firms with Strategic Market Status (SMS) and impose conduct requirements. The CMA publishes market investigations, merger decisions, and enforcement notices — all strong sources for tech journalists. CMA open cases are listed on the gov.uk CMA pages and often generate months of follow-up stories.