About FSM

People first — the movement for disabled-led Physical AI

Our story, our vision, our eight pillars — and how you can be part of the most important systems redesign of our generation.

Part I · The Scale of Crisis

Four interlocking crises that demand a new model

The current care system is failing — not because the people in it aren't trying, but because it was designed for a world with abundant labour, modest demand, and no viable alternative. None of those conditions still hold.

Crisis 1

The Workforce Crisis

Adult social care in England employed 1.52 million people in 2022/23. Vacancy rates consistently exceed 10%. Demand will grow by an estimated 480,000 workers by 2035 — a number that cannot plausibly be recruited, retained, or funded from existing supply.

Crisis 2

The Financial Crisis

Local authority social care budgets exceeded £22 billion in 2023/24 — over 40% of all council expenditure. DWP projects disability benefit expenditure will reach £35 billion per annum by 2028/29. The trajectory is structurally unsustainable.

Crisis 3

The Independence Deficit

More than half of Digital Assistive Technology users say they could not live their lives the way they do without their assistive technology. Yet the current system is built around scheduled carer visits — not the real-time needs of the person.

Crisis 4

The Data Exclusion Problem

Big data techniques systematically fail to represent disabled people, reinforcing biases in favour of non-disabled people. Almost nothing in a service user's home is captured in usable form — deterioration goes undetected until crisis.

Part II · Why the Current Model Fails

The seven failure modes of scheduled care

The national care system was architected around human labour — a care worker visits at scheduled intervals. Between visits, the person is largely unsupported. Seven failure modes compound each other.

1
Temporal coverage gap
20–22 hours of every 24 are unmonitored for most service users. Falls, medication errors, and acute episodes occur in these gaps — unseen and unresponded to.
2
Standardisation over personalisation
Care packages are allocated by assessment category, not by real-time individual need. Formal reassessment is typically annual — a 365-day delay in responding to changing circumstances.
3
Workforce dependency
Every hour of care output requires a corresponding unit of human labour. Any workforce shortage, staff absence, or cost increase immediately reduces care capacity — with no buffer.
4
Lack of continuity
Many service users see different carers at each visit. Relationships — crucial to psychological wellbeing — cannot form under rotating care schedules.
5
Dignity erosion
Physical assistance with bathing, toileting, and mobility requires intimate human presence. Many disabled people experience this as a profound loss of privacy and dignity — not a preference, a reality.
6
Data poverty
Almost nothing in a service user's home is captured in usable form. Deterioration goes undetected until crisis. Prevention is structurally impossible in a system with no continuous data stream.
7
Geographic exclusion
Rural areas cannot recruit care workers. Quality of care depends on where you live. Physical AI infrastructure does not share this dependency — it is geography-neutral by design.
Part III · Our Architecture

PAI + FSM: built together, owned by people

PAI Physical Capability CIC and Future of Smart Machines CIC form an interlocking social enterprise structure — Parliamentary presence and open standards from FSM, deployment and research from PAI.

PAI Physical Capability CIC

Research, Consultancy & Deployment

Operates five programme pillars: Research & Standards, Consultancy (ADVISE / BUILD / TRAIN / OPERATE), Open Ecosystem Programme with BSI, National Care Service Integration, and Community Co-Design. Mission-rate pricing ensures accessibility.

Future of Smart Machines CIC

Parliamentary Presence & Standards

Provides APPG on Robotics secretariat, open standards advocacy, and ecosystem convening. Ensures that policy, regulation, and public procurement of Physical AI are shaped by disabled people — not by vendors.

Part IV · The 10× Transformation

Eight pillars of change — from dependency to independence

Each pillar shows what the current care system delivers — and what Physical AI makes possible. Not incremental. Transformational.

BCG's 2026 analysis argues that AI requires 'a fundamental rethink of care delivery and access.' PAI's thesis goes further: Physical AI does not merely improve the existing model. It enables a different one — one that is categorically superior on every dimension that matters to disabled people.

Pillar 1 · Time & Presence

From 2 hours to 24 / 7

Today
2–4 hours of direct human contact per day. 20+ hours unmonitored — unseen and unprotected.
→
10× State
Continuous presence — monitoring vital signs, detecting falls, prompting medication, providing mobility assistance around the clock.
Target: 24-hour coverage at the cost of a 2-hour human care package — a 10× coverage expansion.
Pillar 2 · Personalisation

From category to continuous individual

Today
Standardised care package assessed by category. Formal reassessment once a year.
→
10× State
Physical AI continuously learns individual patterns, preferences, and needs. Small-data, few-shot models personalise support daily — no assessment cycle needed.
Target: Continuous personalisation updated daily vs. annual reassessment — a 365× improvement in responsiveness.
Pillar 3 · Independence & Agency

From dependency to agency

Today
Service users organise their lives around carer visit times — meals, baths, outings all scheduled around someone else's availability.
→
10× State
Cobotic systems allow a person to bathe when they choose, cook when they want, move freely. Physical AI scales this to everyone.
Target: Disabled person controls when and how assistance is delivered — not when a carer is available.
Pillar 4 · Workforce Leverage

From labour to capital

Today
Every unit of care output requires one unit of human labour. Shortages mean immediate gaps in care.
→
10× State
Human carers shift from routine physical tasks to supervision, relationships, and complex clinical care. One worker supervises 8–12 robot-assisted users vs. 3–4 today.
Target: One care worker supervises 8–12 robot-assisted service users vs. 3–4 today — a 3× productivity improvement.
Pillar 5 · Early Intervention

From crisis to prevention

Today
Deterioration detected only at carer visits. Falls detected when the person cannot get up. Medication missed until the next visit.
→
10× State
Continuous monitoring detects gait changes triggering physiotherapy review. A 20% reduction in falls-related NHS costs would save £400 million annually.
Target: Falls detected in seconds vs. hours. Deterioration identified weeks before crisis — not after hospitalisation.
Pillar 6 · Dignity & Privacy

From intimate dependency to privacy-preserving assistance

Today
Intimate physical assistance from potentially unknown carers. Profound loss of privacy in one's own home.
→
10× State
User-controlled cobotic systems, bathroom assistance robots, and AI-powered mobility aids restore privacy — on the person's own schedule, in their own space.
Target: Service user controls their own care assistance without requiring external human presence for intimate tasks.
Pillar 7 · Geographic Equity

From postcode lottery to universal access

Today
Rural areas cannot recruit care workers. Quality of care depends entirely on where you happen to live.
→
10× State
Physical AI infrastructure does not depend on local labour supply. A care robot works in a rural Welsh farmhouse as effectively as central London. Telehealth integration from anywhere.
Target: Quality-adjusted care access equalised across urban/rural geographies — no postcode lottery.
Pillar 8 · Data & Learning

From data poverty to continuous learning

Today
Almost nothing in the home is captured in usable form. The system cannot learn from outcomes. Prevention is structurally impossible.
→
10× State
Physical AI generates continuous structured data — mobility, sleep, medication, social engagement, vital signs — aggregated with consent and sovereignty protection.
Target: Real-time care outcome data vs. annual assessment snapshot — transforming what commissioners can see and act on.
Part V · Get Involved

How you can shape the Future of Smart Machines

Co-design with us

If you are a disabled person or carer, your lived experience is the most important input we have. Join our co-design panels to shape how Physical AI systems are built, tested, and deployed.

Nothing about us without us. →

Commission research

Are you a local authority, NHS Trust, or government body looking to understand Physical AI in care? PAI CIC offers ADVISE, BUILD, TRAIN and OPERATE service lines at mission rates to public bodies.

Work with PAI CIC →

Partner in the National Care Service

The Labour Government's National Care Service represents the defining deployment opportunity for Physical AI. We are building a coalition of providers, commissioners, and technology partners to shape this window.

Join the NCS coalition →

Follow our work at Parliament

FSM runs the secretariat for the APPG on Robotics. Follow our Parliamentary work, contribute evidence to inquiries, and help ensure that robotics policy in the UK is shaped by those who need it most.

APPG on Robotics →

Ready to co-design the future?

Join our community of disabled people, researchers, and policymakers building this together.

Contact us →