A National Social Infrastructure Plan

Britain does not need a Ministry of Friendship.

Government cannot create intimacy, prescribe sociability or decide how people should spend their private lives.

It can influence whether people have practical routes into shared life.

Schools, workplaces, transport, housing, youth services, libraries, sport, public space and digital systems already shape those routes. The problem is that each is governed for another primary purpose.

A National Social Infrastructure Plan would not replace those purposes. It would require government to account for the contact, participation and capability produced between them.

Its goal would be narrow:

By 2036, every young person should have realistic routes to regular activity, a trusted adult and independent movement; to useful work or contribution; and to human help when digital provision is insufficient.

Policy boundary. Social infrastructure means the conditions that make voluntary contact and participation possible. It does not mean compulsory socialising, surveillance of relationships or treating loneliness as personal failure.

Education, youth services, housing and much of transport are devolved. England could adopt this plan directly in those fields; UK-wide measures would require agreement with Scotland, Wales and Northern Ireland.

Institution One: a Social Infrastructure Mission

The Cabinet Office creates a ten-year mission spanning education, work, health, transport, housing, culture, local government and digital regulation.

A named minister chairs it. An independent commissioner or statutory council publishes annual progress. Local government participates through agreed outcomes rather than uniform delivery; devolved administrations decide whether and how to align.

The mission has three responsibilities:

1. aggregate indicators that departments currently hold separately;
2. test major reforms for lost social function;
3. intervene where several systems fail the same group or place.

It does not run youth clubs or buses from Whitehall. It makes fragmentation visible and assigns accountability.

Instrument One: the Social Interface Impact Note

Major policy and service redesigns answer five questions:

• Which routine human interface changes or disappears?
• Who uses it for more than its formal purpose?
• Is the replacement optional, accessible and reversible?
• How will beginners practise the lost human capability?
• What indicator would reveal unintended withdrawal?

The note applies to significant digitisation, automation, closures, remote-service models and public-realm changes. It is short and published.

It does not create a veto against efficiency. It requires the trade-off to be named.

Guarantee One: a first rung into adult work

Every 16- to 24-year-old outside education, employment or training for four months receives a quality offer: paid work, apprenticeship, supported training or compensated civic service.

The existing Youth Guarantee, foundation apprenticeships and local trailblazers provide the starting machinery.

Quality requires:

• a named supervisor;
• real contact with colleagues, customers or community;
• structured feedback;
• protected learning time;
• a reference or recognised credential;
• progression review after completion.

Government should first pilot sector agreements in which employers reducing substantial junior task bundles support placements, shared academies or a training fund. Any compulsory contribution should follow evidence, not precede it.

The purpose is not to preserve obsolete tasks. It is to preserve the process by which a novice becomes useful.

Guarantee Two: somewhere to go and somebody who stays

The National Youth Strategy’s 2035 ambitions become statutory reporting duties.

Every local area maps:

• regular free or affordable activities;
• trusted-adult capacity;
• evening and weekend opening;
• disability access;
• journey time by public transport;
• participation and dropout by deprivation.

Capital and revenue funding are paired. A building cannot be counted as provision without people, hours and an operating plan.

Trust is measured through young people’s reported relationship and duration, not a completed introduction.

Schools become access points, not sole providers. The Finnish model offers a useful mechanism: enjoyable activity connected to the school day, free at the point of use and chosen with pupils.

Guarantee Three: independent youth mobility

Young people cannot participate in places they cannot reach.

For England, a national framework provides predictable, low-cost bus travel for under-22s, with local enhancement and integration across operators. Scotland’s existing under-22 scheme supplies a practical precedent, not proof of transferability. Funding rewards evening, weekend and cross-boundary access to colleges, work, youth facilities and town centres.

Local Transport Plans report a Youth Access Standard:

• share of young residents reaching education or entry work within a reasonable journey;
• evening access to activity;
• independent school travel;
• rural and disability gaps;
• affordability relative to youth income.

The aim is not free travel everywhere regardless of cost. It is a legible minimum and explicit trade-offs.

Guarantee Four: human education around AI

Every school and college adopts a human-development standard for AI.

AI may support explanation, accessibility, feedback and teacher administration. It must not remove responsibility for reasoning, discussion, practical work, collaboration and oral defence.

Key requirements:

• pupils disclose material AI assistance;
• assessment includes work that demonstrates independent understanding;
• AI literacy includes uncertainty, persuasion, privacy and commercial incentive;
• remote provision has a return-to-participation plan;
• teacher time released by automation is tracked and redirected towards pupils;
• no child is assigned an AI companion as a substitute for pastoral care.

The test is not whether AI is present. It is whether its use increases human capability.

Guarantee Five: a right to human escalation

Digital public services remain available and improve. A practical human route remains for consequential decisions and vulnerable users.

The right is not immediate face-to-face service for every transaction. It is timely escalation where:

• automated advice cannot resolve the issue;
• a decision significantly affects rights, income, health or education;
• disability, language or distress makes the interface unsuitable;
• the person disputes the system's understanding.

Public bodies publish waiting time and resolution data for human escalation. Efficiency savings cannot be measured without the cost of failed access.

Rule One: regulate emotional systems by relationship risk

AI companions should be assessed not only for prohibited content but for relational design.

The regulatory framework should examine:

• claims or simulations of personhood;
• encouragement of exclusivity or dependency;
• sexualised interaction with minors;
• handling of self-harm and crisis disclosures;
• memory and use of intimate data;
• commercial prompts triggered by vulnerability;
• clarity that the user is interacting with a system;
• routes to human support.

Age restrictions may be justified for high-risk functions. The evidence base should be reviewed independently and updated as systems change.

The goal is not to ban benign assistance or adult choice. It is to prevent commercial systems from optimising emotional dependence without accountability.

Rule Two: pair digital restriction with physical opportunity

Every major child-online-safety measure includes an Offline Opportunity Statement.

It identifies what beneficial function may be lost - friendship coordination, identity support, creative participation or community - and how schools, youth services and local organisations will support alternatives.

This does not require government to replicate a platform. It prevents protection policy from assuming that children already possess somewhere else to go.

Investment One: public places that tolerate presence

Local regeneration and planning guidance require provision for teenagers and young adults who are not purchasing, formally supervised or merely passing through.

Libraries, parks, covered seating, courts, youth rooms and civic buildings form a network. Design includes lighting, toilets, weather protection, charging, visibility and reasonable behaviour rules.

Young people participate in design and review.

The standard is not unrestricted use. It is legitimate presence.

Investment Two: housing connected to independence

The ten-year Social and Affordable Homes Programme is assessed partly by whether new homes enable young adults to establish independent lives.

Relevant measures include:

• genuinely affordable rents relative to local earnings;
• secure tenure;
• access to work and public transport;
• availability of smaller homes and shared options;
• proximity to services and community space.

Housing policy cannot be reduced to social connection. It should stop treating location and independence as socially neutral.

Measurement: a dashboard without a single score

The Office for National Statistics and departments publish an annual Social Participation Dashboard.

It includes:

• persistent and severe school absence;
• NEET and economic inactivity, with data-quality warnings;
• under-21 entry into paid work and apprenticeships;
• activity participation and trusted-adult relationships;
• independent travel and accessibility;
• housing transitions;
• in-person contact, perceived support and loneliness as separate measures;
• use of AI for education, work and emotional support;
• availability and performance of human public-service routes.

Results are broken down by age, income, disability, ethnicity, sex and place where sample quality allows.

No composite national score should conceal trade-offs or weak data.

Delivery sequence

2026-2028: establish responsibility

• create the mission and impact note;
• baseline the dashboard;
• standardise Youth Guarantee quality;
• define trusted-adult and participation measures;
• issue AI education and companion safeguards;
• identify transport and provision deserts.

2029-2031: scale what survives evaluation

• expand effective first-rung routes;
• pair youth capital with multi-year revenue;
• establish the youth travel minimum;
• embed human escalation in public-service standards;
• target places where several indicators fail together.

2032-2036: make the system ordinary

• renew or amend the Youth Strategy using observed outcomes;
• place successful guarantees on durable statutory and funding foundations;
• revise AI rules as capability and use change;
• report whether the participation gap and trusted-adult ambitions were achieved in lived experience.

Cost and limits

This plan requires public money, employer participation and local capacity. It will produce deadweight, uneven delivery and programmes that fail.

Transport in rural areas is expensive. Youth workers cannot be trained instantly. Human public services cost more than automated ones. Entry guarantees can become low-quality schemes. Regulation can entrench large firms, invade privacy or displace use. Metrics can distort behaviour.

Government cannot guarantee friendship, confidence or belonging.

The alternative is not cost-free. Persistent absence, economic inactivity, crisis mental-health demand and lost capability also carry fiscal and human costs.

The plan’s claim is limited:

Government should be accountable for whether people can enter shared life, not for what private relationships they form once there.

Evidence used across this analysis: [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12]

Source notes

Evidence cut-off: 17 August 2026. Forecast and scenario limits are identified in the text.

1. DCMS, Youth Matters: Your National Youth Strategy. Open source ↩ Return
2. Department for Work and Pensions, Get Britain Working White Paper. Open source ↩ Return
3. Department for Work and Pensions, Young people and work: interim report. Open source ↩ Return
4. Finnish Ministry of Education and Culture, The Finnish model. Open source ↩ Return
5. Department for Transport, Local Transport Plans guidance (2 April 2026). Open source ↩ Return
6. Ofcom, AI chatbots and online regulation. Open source ↩ Return
7. UK Government, International AI Safety Report 2026. Open source ↩ Return
8. UK Government, Growing up in the online world: progress statement. Open source ↩ Return
9. Ministry of Housing, Communities and Local Government, A decade of renewal for social and affordable housing. Open source ↩ Return
10. OECD, Social Connections and Loneliness in OECD Countries (2025). Open source ↩ Return
11. Cabinet Office, Government of Japan, Measures to Address Loneliness and Isolation. Open source ↩ Return
12. Transport Scotland, Year One Evaluation - Young Persons’ Free Bus Travel Scheme (2023). Open source ↩ Return