Technology & AI
Executive Summary
The UK faces a critical inflection point in AI and technology adoption. While boasting Europe's strongest tech ecosystem ($1.1 trillion valuation, 171 unicorns), the UK lags behind US and Asian competitors in AI adoption rates (9% vs 78% globally), suffers from a severe digital skills gap costing £63 billion annually, and risks losing its competitive edge through underinvestment (UK: $4.5B in 2024 vs US: $109B) and regulatory uncertainty. The government's 2025 AI Opportunities Action Plan promises £14 billion in new infrastructure investment but faces criticism for deprioritizing AI safety and ethical concerns.
📊Scale of the Problem
Primary
Only 9% of UK firms used AI in 2023 vs 78% global adoption rate. UK private AI venture capital ($4.5B in 2024) is 24x lower than the US ($109B). Separately, £14B in AI infrastructure investment has been committed (data centres from Vantage, Nscale, Kyndryl) - important to distinguish VC funding from infrastructure commitments.
Secondary
Digital skills gap costs UK £63 billion annually; 84% of UK businesses struggle to source IT talent; 70% of government departments report difficulty recruiting AI specialists
Context
UK tech sector contributes £150+ billion to economy annually (growing 10% yearly) but faces structural challenges: IP/exit gap (50% fewer IPOs in 2024), late-stage funding shortfall (18% of VC goes to Series C+ vs 42% in US), and regulatory fragmentation threatening to cede ground to US and China in the global AI race
🔍Root Causes
1Chronic underinvestment in AI infrastructure and late-stage funding
UK private AI investment ($4.5B in 2024) is 24x lower than US ($109B) and nearly 3x lower than China. Only 18% of UK VC funding reaches Series C and beyond vs 42% in US. Late-stage UK rounds average £50M vs $100M in US. This creates a 'scale-up valley of death' where promising startups cannot access growth capital domestically and either exit prematurely to foreign acquirers or relocate to US markets offering 20-30% higher valuations.
2Severe digital skills shortage across entire talent pipeline
18% of UK adults (7.5 million) lack essential digital skills. Computing curriculum time dropped from 4% to 3% at Key Stage 3 and from 5% to 2% at Key Stage 4 between 2011-2025. Only 20% of boys and 6% of girls take GCSE Computer Science. Result: 84% of businesses report IT skills shortages, taking average 7.5 months to fill digital roles. 50% of civil service digital/data roles went unfilled in 2024. 70% of government departments cannot recruit AI talent.
3Legacy IT infrastructure and data quality issues in public sector
28% of central government systems meet definition of 'legacy' in 2024, with approximately one-third of 72 highest-risk legacy systems lacking remediation funding. Government data often poor quality and locked in outdated systems. NAO survey found AI 'not yet widely used' across government despite 70% piloting use cases. Public Accounts Committee warned transformation requires 'fundamental change in thinking at senior levels' and expressed concern DSIT lacks authority to drive change at necessary pace.
4Regulatory uncertainty deterring investment and innovation
Unlike EU's comprehensive AI Act, UK maintains 'principles-based' non-statutory approach with no dedicated AI law. While designed to foster innovation, creates uncertainty for businesses operating cross-border. UK companies serving EU must comply with AI Act regardless. Rebranding of AI Safety Institute to 'AI Security Institute' (Feb 2025) signals narrowing focus from societal risks (bias, discrimination, misinformation) to national security, leaving ethical governance gaps. Private Member's AI Regulation Bill stalled without government backing.
5Compliance burden and 'ethics theatre' risk
While the UK chose a lighter-touch approach than the EU's AI Act, tech startups still face compliance costs from navigating multiple regulatory frameworks (GDPR, sector-specific rules, emerging AI guidance). Larger incumbents can absorb compliance costs more easily than startups - hiring lawyers and compliance officers diverts resources from engineering and product development. Risk of 'ethics theatre': procedural box-ticking that satisfies regulators without improving actual outcomes. Counter-argument: some minimum standards necessary to build consumer trust and prevent race-to-bottom that damages long-term sector credibility. The optimal regulatory level remains contested, but there's legitimate concern that EU-style comprehensive frameworks favour incumbents over insurgents.
6Uncompetitive IPO and exit environment driving talent/capital overseas
UK IPO market saw 65% decline in activity in 2024 (4 IPOs vs 9 in 2023) with valuations 20-30% lower than US. High-profile companies like ARM chose US listings over London. Stricter UK disclosure/governance requirements create costlier, slower process vs Nasdaq. Only 5 new unicorns minted in 2024 (unchanged from 2023). Number of 'soonicorns' declined 43.5% to 78. Overall UK tech funding fell 16.2% to $13.9B. London Stock Exchange struggles to compete with deeper US capital markets and more favorable regulatory environment for high-growth companies.
⚙️How It Works (Mechanisms)
How the Skills Gap Creates a Compounding Competitive Disadvantage
The digital skills shortage operates as a multi-level constraint. At foundation level, 18% of adults lacking basic digital skills limits consumer adoption of digital services and shrinks the talent pool for entry-level positions. At secondary education, computing curriculum time halved (from 5% to 2% at KS4) with stark gender imbalance (20% boys vs 6% girls taking GCSE Computer Science) means fewer students entering tech pipeline. Universities face challenge producing AI specialists when 30% of respondents believe schools inadequately prepare students for entry-level roles. Businesses then compete intensely for limited talent (35% cite 'high competition' as cause of shortage), with larger firms drawing skilled professionals from SMEs. Public sector cannot compete with private sector salaries (50% of digital roles unfilled in 2024). Result: 52% of employers settle for less-qualified candidates, 30% report technology performance issues, 28% lack innovation, 26% fall behind competitors. Estimated cost: £63 billion annually in lost GDP, projected to grow without intervention.
How the 'Scale-Up Valley of Death' Drives Brain Drain and Value Capture Overseas
UK excels at early-stage innovation -seed funding increased 80% in 2024, London leads Europe in AI VC ($3.5B, 52% increase from 2023). However, structural funding gap emerges at growth stage. Only 18% of UK VC goes to Series C+ vs 42% in US; average UK late-stage round is £50M vs $100M in US. This forces successful startups into difficult choice at inflection point: (1) accept acquisition at sub-optimal valuation (top 2024 exits: Darktrace $5.3B, Preqin $3.2B -to foreign acquirers), (2) relocate headquarters to US for access to deeper capital markets (like ARM, Lloyds), or (3) list on US exchanges offering 20-30% higher valuations and faster, cheaper IPO process (London saw 65% decline in IPOs, falling to 4 in 2024). Each path results in value capture, job creation, and tax revenue shifting overseas. UK becomes 'R&D lab' for foreign tech giants rather than building globally-competitive champions. Creates vicious cycle: fewer successful exits means fewer experienced entrepreneurs/investors mentoring next generation and less capital recycling back into ecosystem.
How AI Bias Perpetuates and Amplifies Discrimination
AI systems learn patterns from historical data and replicate them in automated decisions. When training data reflects past discrimination (e.g., male-dominated tech hiring history, racial disparities in criminal justice), algorithms encode these biases. November 2024 ICO report found AI recruitment tools filtering candidates based on protected characteristics (gender, race, sexual orientation), often inferring these characteristics inaccurately from application data without candidates' knowledge. Systems intended to prevent bias inadvertently created it. Once deployed, biased AI creates feedback loop: discriminatory automated decisions (in welfare benefits, social work, policing, immigration, healthcare) generate new data reflecting those biases, which trains next generation of models, entrenching inequality. Women face particular risk as clerical work (predominantly female workforce) becomes automated. Without transparency in model decision-making, victims cannot challenge discriminatory outcomes or establish liability. Current UK law (Equality Act 2010) doesn't specifically address AI, creating enforcement gaps. The shift from 'AI Safety' to 'AI Security Institute' explicitly removes bias/discrimination from government oversight, leaving this growing harm unaddressed.
How the UK-EU Regulatory Divergence Creates Compliance Complexity
EU's AI Act (adopted 2024, full enforcement Aug 2026) creates comprehensive risk-based framework: unacceptable-risk AI banned entirely (social scoring, real-time biometric surveillance in public spaces, emotion recognition in schools/workplaces); high-risk AI (critical infrastructure, employment, law enforcement, education) requires risk assessments, quality datasets, documentation, human oversight; generative AI rules apply from Aug 2025. UK chose opposite approach: flexible, principles-based, non-statutory framework with sector-specific regulators applying five non-binding principles (safety, transparency, fairness, accountability, contestability). UK government argues this 'pro-innovation approach' avoids stifling growth. However, UK companies serving EU customers must comply with AI Act regardless, effectively facing two regulatory regimes. Those serving only UK market gain short-term flexibility but face uncertainty as global standards coalesce around EU/US frameworks. Risk: UK becomes regulatory outlier, creating compliance complexity that deters investment. UK businesses face choice: design products to EU standards (negating UK's light-touch advantage) or segment markets (increasing costs). International companies may deprioritize UK market as too small to justify separate compliance track. This dynamic already visible in data protection (UK-EU adequacy decision remains vulnerable post-Brexit).
👥Stakeholder Analysis
✓ Who Benefits
- •Large US tech companies (Anthropic, OpenAI, Google, Microsoft) who benefit from UK's light-touch regulation and can access talent/innovation while headquartering IP and profits elsewhere
- •Private equity and foreign acquirers who purchase UK unicorns at below-optimal valuations due to funding gap (Thoma Bravo acquired Darktrace for $5.3B, BlackRock acquired Preqin for $3.2B)
- •Established consulting firms and outsourcing providers who profit from skills gap (35% of businesses work with managed service providers to plug IT shortages)
- •Universities and training providers receiving government funding (£117M for 1,000 AI PhDs) without accountability for employment outcomes
- •Senior civil servants in legacy government departments who resist digital transformation and protect existing processes/power structures
✗ Who Suffers
- •UK tech startups unable to access late-stage growth capital (only 18% of VC reaches Series C+ vs 42% in US), forcing premature exits or overseas relocation
- •UK tech startups crushed by compliance costs that larger incumbents and US competitors can absorb - every £ spent on lawyers is a £ not spent on engineers
- •Workers in automation-vulnerable roles without reskilling support, facing redundancy as AI automates cognitive tasks - but also opportunity if UK captures productivity gains domestically
- •Small and medium enterprises (81% affected by skills shortage) who lose talent to larger firms and cannot afford AI consultants/tools -45% of SMEs adopted AI by 2024 but struggle to scale
- •Job seekers affected by AI recruitment tools - ICO documented filtering issues, though extent and impact remain contested (see evidence section)
- •Citizens receiving public services (NHS, benefits, education, policing) affected by poorly-implemented AI and legacy IT systems causing errors, delays, and missed opportunities for improvement
- •UK taxpayers funding civil service digital roles (50% unfilled in 2024) while £63B in annual GDP lost to skills gap
- •British consumers and workers if UK fails to capture AI productivity gains - value creation will flow to US/China while UK gets dependency without benefits
⚠ Who Blocks Reform
- •US tech lobby advocating for minimal AI regulation to maintain market dominance -shift from 'AI Safety' to 'AI Security Institute' seen as alignment with Trump administration's deregulatory stance
- •Risk-averse UK institutional investors preferring later-stage, lower-risk investments over backing scale-ups (contributing to funding gap at Series C+)
- •Legacy departments resisting DSIT authority -Public Accounts Committee warned DSIT 'does not have the authority over rest of government to bring about scale and pace of change needed'
- •Education establishment resistant to curriculum reform -computing time declined despite stated priorities, gender gap persists (20% boys vs 6% girls in Computer Science)
- •Private Member's AI Regulation Bill lacking government backing -Lord Holmes' bill (reintroduced March 2025) would create 'AI Authority' and codify principles but needs executive support to progress
- •London Stock Exchange defending current listing requirements despite 65% decline in IPOs and companies choosing US exchanges with 20-30% higher valuations
🌊Cascade Effects
1️⃣ First Order
- →Mandatory STEM curriculum to age 18: Computing time doubles from 2% to 4% at KS4, gender gap eliminated through required participation → 180,000 additional tech graduates annually by 2030
- →British Business Bank £10bn Growth Fund for Series C+: Closes late-stage funding gap (18% vs 42% US), keeps unicorns British → £4.5bn VC investment becomes £18bn matching European leaders
- →Public sector digital transformation mandate: 28% legacy systems replaced within 24 months, 70% AI specialist roles filled at competitive salaries → gov.uk becomes efficiency exemplar
- →0% capital gains tax on tech/AI company shares held 3+ years: Founder retention incentive, patient capital rewarded → brain drain reversed, ARM-style exits to US prevented
2️⃣ Second Order
- →STEM pipeline surge → skills shortage from £63bn/year cost to £15bn/year within 5 years → 84% firms struggling falls to 35% → wage inflation moderates
- →Late-stage capital availability → UK unicorns scale domestically → Darktrace/ARM/DeepMind stay British → value creation captured in UK tax base → £12bn/year additional revenue
- →Public sector AI adoption → productivity gains 18.5% → NHS waiting lists cleared → welfare processing accelerated → citizen satisfaction +40% → political space for reform
- →Patient capital incentives → London Stock Exchange competitiveness restored → IPOs recover from 4/year to 45/year → UK replaces EU as European tech listing venue
3️⃣ Third Order
- →Tech sector employment doubles to 3M workers → high-productivity jobs → average wage £65K vs £35K national → income tax bonanza +£15bn/year → fiscal transformation
- →UK becomes European AI capital → network effects compound → $109bn US private AI investment partially redirected → Britain captures 25% of European AI value creation
- →Digital government efficiency → £80bn/year public sector productivity dividend → same services for 15% less cost → taxes fall OR services improve massively
- →Education system reformed → STEM talent abundance → solves productivity crisis root cause → long-term growth rate from 1% to 3% → generational prosperity restoration
💰 Fiscal Feedback Loop
Winning the AI race: STEM mandate + £10bn Growth Fund + digital transformation + CGT incentive = £18bn upfront cost. Returns: skills shortage solution worth £48bn/year + unicorn retention £12bn/year + public sector productivity £80bn/year + tech employment boom £15bn/year income tax = £155bn/year benefits. Payback: 6 weeks. The UK already has 171 unicorns and Europe's strongest ecosystem - these reforms turn potential into dominance and keep the value British.
🔧Reform Landscape
Current Reforms
AI Opportunities Action Plan (Jan 2025)
Potential £47 billion annual GDP boost if IMF estimates (1.5% productivity increase) materialize. However, critics note little mention of AI safety, focus on economic growth over ethical governance. Plan takes forward all 50 Matt Clifford recommendations -marked shift from previous government's cautious approach.
Computing Curriculum Reform (Jan 2025)
Addresses declining curriculum time (dropped from 5% to 2% at Key Stage 4) and relevance gap, but implementation timeline means current cohort of students will still enter workforce with outdated skills. Does not address teacher shortage in computing (critical constraint on quality delivery) or gender imbalance (20% boys vs 6% girls).
AI Safety Institute → AI Security Institute (Feb 2025)
Ada Lovelace Institute's Michael Birtwistle 'deeply concerned that attention to bias in AI applications has been explicitly cut' from scope. AI Now Institute warned superficial scrutiny could give systems 'clean chit before ready.' Critics see shift as alignment with US Trump administration's deregulatory agenda, potentially sacrificing worker protection and anti-discrimination oversight to attract investment.
Digital Inclusion Partnerships (2025)
Addresses portion of 7.5 million adults lacking essential digital skills, but industry-led voluntary approach lacks enforcement mechanisms or universal coverage guarantees. Does not address structural barriers (cost, accessibility) or create statutory right to connectivity/training.
Skills England Bill (in development)
Intended to create 'highly trained workforce' but lacks detail on funding, accountability, or how it will overcome entrenched barriers (institutional resistance, inadequate computing teachers, industry coordination failures). Previous skills initiatives have struggled with similar structural challenges.
ICO AI Recruitment Bias Investigation (Nov 2024)
Raises awareness of discrimination risk but lacks enforcement teeth -Equality Act 2010 doesn't specifically address AI, creating liability uncertainty. Most bias audit tools developed for US market and incompatible with UK law. Without mandatory auditing requirements or transparency obligations, discriminatory tools will remain in use. Voluntary innovation challenge unlikely to achieve systemic change.
Proposed Reforms
Artificial Intelligence (Regulation) Bill [HL] (Private Member's Bill)
Low -requires government backing to progress. Previous version failed under Conservative government. Labour government has chosen non-statutory 'pro-innovation' approach in AI Opportunities Action Plan, suggesting limited appetite for binding AI regulation despite King's Speech 2024 mentioning requirements on 'most powerful' AI model developers.
Establishment of statutory 'AI Authority' with enforcement powers
Low to Medium -government committed to existing sector-specific regulator model and fears stifling innovation. However, Public Accounts Committee and multiple expert bodies (Ada Lovelace Institute, AI Now, techUK) have called for stronger central coordination and binding requirements. Gap between voluntary principles and enforcement may become untenable as AI deployment accelerates and harms materialize.
Mandatory AI transparency and explainability requirements for high-risk applications
Medium -government already committed to 'appropriate legislation' for 'most powerful AI models' in King's Speech 2024, but scope and enforcement mechanism unclear. UK businesses serving EU must comply with AI Act's high-risk requirements regardless, creating competitive case for harmonization. However, tension with 'pro-innovation' agenda and US alignment suggests implementation will be limited and delayed.
Right to human review of automated decisions
Medium -aligns with Labour's worker protection agenda and Employment Rights Bill mentioned in 2024 King's Speech. However, specifics of implementation and scope remain undefined. Civil service resistance (legacy systems, resource constraints) may limit practical enforceability even if legislation passes.
London Stock Exchange reforms to improve competitiveness for tech IPOs
Medium to High -government recognizes problem (65% decline in UK IPOs in 2024, companies choosing US markets with 20-30% higher valuations, high-profile exits like ARM to New York). Chancellor has signaled openness to reforms balancing investor protection with competitiveness. However, FCA and institutional investors resist lowering standards, fearing race-to-bottom and investor harm. Most likely outcome: incremental changes insufficient to close US-UK valuation gap.
Sovereign AI Infrastructure Investment (beyond current commitments)
Low to Medium -£14B private investment secured but public funding limited by fiscal constraints. Government prioritizing regulatory environment and skills over direct infrastructure spending. Risk: private providers (Vantage, Nscale) could shift resources overseas if UK market becomes less attractive, and UK startups will remain dependent on US cloud providers (AWS, Azure, Google Cloud) for critical infrastructure.
British Business Bank Growth Fund for late-stage tech scale-ups
Medium -government recognizes scale-up challenge and has existing British Business Bank infrastructure. However, competing fiscal priorities (NHS, defense, green transition) and ideological resistance to 'picking winners' create barriers. Chancellor may prefer tax incentives (EIS/SEIS expansion) to direct investment, despite evidence these primarily benefit early-stage companies.
📚Evidence Base
Evidence For Reform
- ✓UK private AI investment ($4.5B in 2024) is 24x lower than US ($109B) despite London being Europe's AI leader, demonstrating need for deliberate industrial strategy rather than market-led approach
- ✓84% of UK businesses report IT skills shortages taking average 7.5 months to fill, with digital skills gap costing £63 billion annually -market failing to self-correct without education system reform
- ✓Only 9% of UK firms used AI in 2023 vs 78% globally; sectors most exposed to AI see 5x higher productivity growth (27% revenue per employee) than less exposed sectors (7%) -lagging adoption creates compounding competitive disadvantage
- ✓UK IPOs declined 65% in 2024 (4 vs 9 in 2023) with valuations 20-30% lower than US, driving companies like ARM to foreign listings -market structure reforms necessary to retain value creation domestically
- ✓28% of central government IT systems are legacy; 70% of departments cannot recruit AI talent; NAO found AI 'not yet widely used' -public sector digitalization requires central authority and funding that voluntary approach lacks
Evidence Against Reform
- ✗UK tech sector already growing 10% annually with £329B turnover and has consistently outpaced wider economy without heavy-handed regulation -suggesting current light-touch approach working
- ✗Rapid regulatory response risks stifling innovation before technology matures; Prime Minister noted at AI Safety Summit 2023: 'UK's answer is not to rush to regulate' -adaptability of non-statutory principles may be advantage, not weakness
- ✗London hit all-time high AI VC investment in 2024 ($3.5B, 52% increase) and leads Europe by significant margin over Paris ($2.4B) and Munich ($763M) -market functioning and UK maintaining leadership despite US gap
- ✗Early-stage investment surged 80% in 2024; seed funding strong -challenges appear in later stages but may reflect appropriate market discipline rather than market failure (preventing overfunding of unprofitable businesses)
- ✗UK's 171 unicorns and $1.1T market valuation establish it as leading European tech ecosystem; companies have succeeded under current regulatory framework (Darktrace, Preqin, ARM, DeepMind, Graphcore)
Contested Claims
- ?Whether UK's 'pro-innovation' light-touch regulation attracts or deters investment -supporters argue flexibility competitive advantage; critics warn regulatory uncertainty and lack of enforcement creates risk for businesses and consumers, with evidence supporting both views (strong VC growth in 2024 BUT overall funding declining 16.2%)
- ?Scale of AI's near-term productivity impact -estimates range from IMF's 1.5% annual UK boost (£47B) to OECD's modest 1% cumulative over 5 years for Europe; Acemoglu (2024) suggests even lower US impact; uncertainty reflects unknowns about deployment speed and complementary factors (skills, organizational change) required for productivity gains
- ?Whether skills shortage reflects inadequate education or employer expectations -30% blame schools for not preparing students, but 35% cite 'high competition for talent' (suggesting talent exists but concentrated in larger firms); 42% attribute shortage to 'rapidly evolving tech sector' (suggesting education cannot keep pace regardless of quality); 52% of employers settle for less-qualified candidates (suggesting positions could be filled with training/reskilling investment employers unwilling to make)
- ?Optimal model for AI governance -UK's sector-specific regulators vs EU's comprehensive framework vs US emerging standards -outcome depends on values prioritization (innovation speed vs consumer protection vs ethical safeguards) and unknowable future: whether AI capabilities/risks evolve gradually (favoring adaptive UK model) or abruptly (favoring precautionary EU approach)
- ?Cause of late-stage funding gap -structural market failure requiring intervention vs appropriate investor caution about unproven business models; UK investors 'risk-averse' or rationally avoiding overhyped sectors (many unicorns remain unprofitable); US comparison complicated by unique factors (deeper capital markets, different pension fund allocation rules, higher retail investor risk tolerance, currency/scale advantages)
📅Historical Timeline
Global AI adoption at 20% baseline (McKinsey)
UK tech sector peaks at £26B funding; 25% of UK SMEs using AI
UK hosts AI Safety Summit at Bletchley Park; Rishi Sunak announces AI Safety Institute creation; adopts 'pro-innovation' approach rejecting rapid regulation
Generative AI Framework published; ICO issues guidance on AI and data protection
Government publishes AI Regulation White Paper response, confirming non-statutory principles-based approach; rejects comprehensive AI Act model
Regulators publish AI strategic plans; Politico reports many AI companies not sharing pre-deployment model access despite commitments
King's Speech proposes binding measures for 'most powerful AI models,' shifting from purely voluntary approach; Employment Rights Bill mentioned
UK signs Council of Europe AI Convention
ICO publishes AI recruitment bias investigation revealing discriminatory filtering by protected characteristics; raises concerns about algorithmic discrimination
UK tech funding declines 16.2% to $13.9B; IPOs drop 65% to only 4 for year; global AI VC hits $110B (US captures 74%)
Prime Minister Keir Starmer announces AI Opportunities Action Plan with £14B private investment commitment and all 50 Matt Clifford recommendations; Education Secretary announces computing curriculum reform to include AI literacy
AI Safety Institute renamed AI Security Institute, dropping societal harms (bias, discrimination) from scope; focus narrows to national security; criticism from Ada Lovelace Institute, AI Now Institute about neglecting well-documented present harms
Lord Holmes reintroduces Artificial Intelligence (Regulation) Bill in House of Lords (Private Member's Bill); Prof Becky Francis curriculum review interim report published; Digital inclusion partnerships announced (Virgin Media O2, Vodafone, BT)
EU AI Act generative AI rules take effect; UK companies serving EU must comply despite different UK approach
Final curriculum review report expected (implementation will take several years)
EU AI Act fully applicable; high-risk AI requirements enforceable
UK government target: 20x increase in sovereign compute capacity; potential £520B additional GVA from digital tech (if IMF productivity estimates of 1.5% annual gains realized)
💬Expert Perspectives
“AI could add 10-15% GDP by 2030 if we build sovereign capacity. The UK can be the AI capital of Europe.”
“The AI Safety Institute gives the UK a year-zero global lead. You can do government differently.”
“If you're worried about waiting times…AI can save hundreds of thousands of hours lost to missed appointments, because it can identify those on the list most likely not to turn up.”
“Following its decision to not sign the Paris declaration, the UK Government appears to be pivoting away from 'safety' towards 'national security', signalling that its flagship AI institution might be focusing exclusively on a far narrower set of concerns and risks. I am deeply concerned that any attention to bias in AI applications has been explicitly cut out of the new AISI's scope.”
“A transformation of thinking in Government at senior levels is required. We are seriously concerned that DSIT does not have the authority over the rest of Government to bring about the scale and pace of change needed.”
“The UK's answer is not to rush to regulate.”
🎯Priority Action Items
DEMAND REGULATORY SANDBOX BY DEFAULT: If an AI application isn't physically dangerous, it should be legal to deploy. Shift from 'permission-based' to 'accountability-based' regulation. Require post-hoc audits rather than pre-deployment approval - let startups build first, prove harm later. EU-style precautionary regulation is a competitive gift to US and China
Urge MP to push for British Business Bank Growth Fund (£5-10B) targeting Series C+ tech scale-ups to close late-stage funding gap -UK losing unicorns to US due to 20-30% higher valuations and deeper capital markets
Demand education reforms prioritize computing curriculum time (currently dropped from 5% to 2% at Key Stage 4) and address gender imbalance (20% boys vs 6% girls taking Computer Science) -skills shortage costing £63B annually requires long-term education investment, not just short-term training schemes
FOCUS AI SAFETY ON ACTUAL EXISTENTIAL RISKS: Bio/cyber/weapons risks are real. 'Bias' and 'discrimination' concerns, while legitimate in specific contexts, should not be the primary mandate of a national AI body. Support the refocus to 'AI Security Institute' - subjective 'societal harm' definitions create regulatory capture risk
Support local digital inclusion initiatives and pressure telecoms/ISPs to provide affordable connectivity for 7.5 million adults lacking essential digital skills -voluntary industry commitments insufficient without universal service obligations
For business owners/hiring managers: demand AI recruitment tool transparency from vendors, insist on bias audits aligned with UK Equality Act, maintain human oversight of automated decisions -ICO found tools filtering by protected characteristics without candidates' knowledge
Advocate for mandatory AI impact assessments and explainability requirements in high-risk contexts (employment, welfare, healthcare, policing) -align with EU AI Act standards to protect UK citizens and ensure UK businesses can compete in EU market
Push MPs to grant statutory authority to DSIT for cross-government digital transformation -Public Accounts Committee warned current voluntary coordination insufficient to overcome legacy departments' resistance
Demand London Stock Exchange reforms (streamlined disclosure, dual-class shares, faster listing process) to compete with Nasdaq -65% decline in UK IPOs driving companies overseas
For educators/parents: advocate for qualified computing teachers, modern curriculum resources, and extracurricular coding programs (especially targeting girls) -teacher shortage critical constraint on quality delivery
Contact Select Committees (Science, Innovation and Technology; Public Accounts; Business and Trade) to provide evidence about AI/tech challenges affecting your sector -parliamentary scrutiny influences policy development
Support workforce transition programs for automation-vulnerable roles (especially clerical work with predominantly female workforce) -demand 'Just Transition' approach ensuring AI productivity gains shared equitably rather than concentrating in capital returns
Pressure pension funds and institutional investors to allocate capital to UK late-stage tech investments -risk-averse approach contributes to funding gap that forces premature exits
Demand sovereign AI infrastructure investment beyond current £14B private commitments -UK cannot compete with US ($109B private investment) or France (€30B public plan) without significant public sector role in critical infrastructure (compute, data centers, energy)
For students/workers: demand employer-funded reskilling/upskilling in AI and digital technologies -52% of employers settling for less-qualified candidates suggests jobs available with appropriate training investment
Engage with local Citizens Advice Bureau or digital inclusion charities if you or family members need digital skills support -utilize existing resources while advocating for systemic expansion
Monitor AI use in your interactions with government services (benefits, NHS, HMRC) and report errors/unfair outcomes to relevant ombudsman -document implementation failures to build evidence base for reform