SHARE.

An emerging Physical AI research consortium

Engineering for Human Value.

Shared Human Advancement through Physical AI Research and Engineering.

Effective Altruism. Applied to Physical AI.

Directing physical intelligence toward the greatest positive impact we can achieve.

Illustrative everyday independence · AI-generated
GUIDED
BY
EFFECTIVE ALTRUISM
FOCUSED
ON
PHYSICAL AI
FOR
SHARED
HUMAN ADVANCEMENT
THROUGHRESEARCH & ENGINEERING

Effective Altruism.
Applied to Physical AI.

Shared Human Advancementthrough Physical AIResearch and Engineering.

Our purpose is human advancement. Our field is Physical AI. We bring leadership, research, engineering and lived experience together to turn that purpose into practical work.

S
Shared
H
Human
A
Advancement
R
Research
E
Engineering

Compassion sets
the purpose.
Evidence guides
the work.

SHARE applies the principles of effective altruism to Physical AI: identify important, neglected needs, compare practical alternatives, and test where research and engineering can make the greatest positive difference.

Our focus is the complete human outcome: wellbeing, agency, opportunity and the distribution of benefits and burdens—today and over time.

Explore the principles of effective altruism
01

Prioritise
neglected needs.

Direct effort toward important problems where an additional contribution can make a meaningful difference.

02

Compare
the alternatives.

Consider the best practical approach, including simpler solutions. Account for costs, constraints and what else those resources could achieve.

03

Test
the human benefit.

Examine who gains, who carries new burdens, what remains uncertain and what evidence should change our direction.

Illustrative scene of two growers examining produce beside an agricultural robot in a sunlit greenhouse.

Livelihoods & shared access

More possibility.
More people.

How could Physical AI support livelihoods, strengthen local capability and widen access to essential resources?

Illustrative concept · AI-generated imagery

For those who can make progress possible

Bring your influence.
Advance human potential.

Leadership can open a working environment, commit engineering capacity, mobilise resources and help useful findings reach an industry.

We invite executives, board members, investors, founders and technology leaders to help shape SHARE and make its first research programmes possible.

Explore leadership & partnership

A proposed Founding Council, connected to research teams and the people whose lives and work Physical AI affects.

Questions worth
working on.

Study the relationship between intelligent machines, the systems around them and the lives they affect.

01

Human benefit

What improves in people's lives—and compared with what?

The investigation

Compare a defined Physical AI application with the best practical alternative. Examine the scale and duration of benefit, full resource costs, uncertainty and potential harm.

What progress could look like

A transparent evaluation method connecting technical performance with wellbeing, time, independence and access, with assumptions available for scrutiny.

02

Agency and work

How can capability grow while people retain meaningful control?

The investigation

Study changes in physical demands, judgment, authority, skills and the organisation of work. Include the people who perform, maintain and supervise the activity.

What progress could look like

Field studies showing where human–robot work expands capability and where it introduces dependency, hidden effort or reduced discretion.

03

Deployment reality

What does reliable performance require from the whole system?

The investigation

Examine models, machines, sensing, environments, safety, maintenance and recovery together. Make human interventions, workarounds and operating limits visible.

What progress could look like

A reproducible work evidence record connecting completion, reliability and recovery with human support, cost and the conditions of deployment.

04

Access and distribution

Who can participate, who gains and who carries the costs?

The investigation

Study affordability, accessibility, ownership and local support. Consider people with limited influence over how technology is chosen and deployed.

What progress could look like

Comparative studies of deployment choices that show how benefits and burdens are distributed across workers, users, communities and organisations.

05

New possibilities

What valuable activity can physical intelligence make possible?

The investigation

Identify useful capabilities and services that are currently unavailable, impractical or inaccessible. Start with a human need and investigate the complete system required to address it.

What progress could look like

Testable opportunity briefs, grounded in lived needs, with a credible path to benefit and clear reasons to continue, adapt or stop.

06

Shared learning

How can knowledge travel across machines, places and people?

The investigation

Study how data, context, memory, skills and failure evidence can be reused responsibly. Identify the assumptions and conditions that limit transfer.

What progress could look like

Open reporting formats and reproducible studies that help others build on findings, including unsuccessful approaches and the boundaries of what works.

Illustrative scene of a skilled technician inspecting a metal component supported by a robotic arm in a workshop.

Human capability & physical work

Support the person.
Improve the work.

Where can physical intelligence reduce demanding effort while preserving human skill, judgment and agency?

Illustrative concept · AI-generated imagery

From a successful demonstration to a useful system

The Real Work
Evidence Project.

What does the complete system accomplish, what does it still require from people, and what new value becomes possible?

We propose a shared method for documenting Physical AI in actual working conditions, connecting technical performance with the human and economic conditions behind it.

Observe
The whole activity: completion, intervention, recovery and maintenance.
Compare
The practical alternative, full costs and quality of the human outcome.
Share
Methods, conditions, limitations and findings others can use.
Help make this study possible

Shared purpose.
Individual contributions.

People, companies and organisations can contribute to SHARE. Each profile connects a person or team with their role in SHARE, their commitment and the work they help make possible.

01 / Individuals

A person. A perspective.

Leaders, researchers, practitioners and people with lived experience, with their capacity and contributions made clear.

02 / Companies

A company. Its contributors.

Participating companies, their contribution to SHARE and the individual people contributing through them.

03 / Organisations

An organisation. Its people.

Research institutions, nonprofits and other organisations, with their focus and named contributors.

Active participants

Loading participant profiles…

The founding directory

Make your contribution part of the story.

Confirmed profiles will show each participant’s role in SHARE, their commitments, research focus and named contributors.

Apply to join SHARE

A shared future
takes shared effort.

Leadership can unlock resources. Research and engineering can test possibilities. Workers, users and communities can show what matters in practice. SHARE needs these perspectives together.

Leadership & partnership

Make the work possible.

Help shape priorities, open a field setting, contribute a capability or resource an independent study.

Explore your leadership contribution

Research & lived experience

Bring what you know.

Bring a question, investigate a claim or share the realities of the work and the people affected.

Prepare a research contribution

Start with the need.

Describe whose life or work could improve, what is missing today, and why the question deserves attention. Connect it to a research area above.

A personal preparation outline. Downloading does not submit an application.

Proposed programme / Research outline

The Real Work Evidence Project

Develop and test a consistent way to describe what a deployed Physical AI system achieves and the resources it needs to do so.

Proposed work record

Initial research approach

Co-design the record with practitioners and affected people, trial it across a small set of contrasting activities, and publish revisions alongside the findings.

This is a proposed research programme. Studies, partners and results will be identified as they are established.