Membership — PASC

About you

Full name

As it should appear on the PASC website

Work email

Affiliation

e.g. "UC San Diego"

Title / role

e.g. "Professor", "Research Scientist".

LinkedIn or personal pageoptional

You are joining as

An individual (researcher, faculty, engineer)Representing my organizationBoth individual and organization

Organization type

Your photo

Shown on the People section if your application is approved. Square crop — upload or paste an image.

Upload photo…

Organization icon

Your organization's logo or icon (PNG/SVG preferred). Shown with your card on the People section — upload or paste an image.

Upload icon…

Participation

How would you like to participate?

Membership and paper collaboration are independent — paper authorship never implies a public membership listing.

Consortium memberBe part of the group defining physical AI safety: listed as a member (with your consent), quarterly updates, feedback on the standard and benchmark. No paper commitment.Paper collaboratorCo-author PASC research — position paper, benchmark, or evaluations. You appear on the papers you author, not in the member list.Paper collaborator + consortium memberBoth of the above.Sponsor or funding partnerSupport shared research, evaluations, workshops, infrastructure, or access without requiring paper authorship.

Research

Which problems do you want to work on?

Read the full problem statements in the Research Agenda →

01 / SpecificationDefine harm precisely enough to evaluate and enforce without losing context.02 / SufficiencyDetermine what bounded memory and evidence are enough for a safety judgment.03 / AnticipationPredict unsafe consequences early enough for an effective intervention.04 / FalsificationFind failures beyond finite benchmarks and clarify what passing can prove.05 / RecoverabilityIdentify when a system still has a feasible path away from harm.06 / InvarianceTest which guarantees survive new hardware, embodiments, environments, and scale.07 / AdversaryTest whether safety mechanisms survive an adaptive opponent inside the loop.

Programs you are interested inoptional

Position paperPhysical AI Safety workshop

Anything else · optional

Anything else?optional