Industry Pain Points

Industry PET failures are rarely about theory alone. They come from procurement, unclear claims, infrastructure gaps, cost surprises, and outputs that still leak.

Procurement-Ready Privacy Claims

Field	Card
Problem	How can PET vendors and builders make privacy claims that procurement teams can evaluate?
The itch	Buyers receive claims like "data stays private" without threat models, assumptions, failure modes, or evidence.
Why it matters	Procurement, legal, and security teams either block useful tools or approve tools they do not understand.
Current workaround	Long questionnaires, one-off security calls, and contract language.
Why the workaround is insufficient	It does not produce comparable evidence across PETs.
What good progress would look like	A claim register format for PET products: protected assets, adversaries, guarantees, evidence, limitations, and required customer controls.
Difficulty	Medium
Good for	Privacy engineer, policy researcher, technical writer
Related PETs	All PETs
Possible first contribution	Draft a one-page procurement evidence form and test it against three PET product claims.

Operational Readiness For PET Deployments

Field	Card
Problem	What operational checks should happen before a PET deployment goes live?
The itch	PET projects often prove the privacy mechanism but under-specify keys, logs, monitoring, incident response, upgrades, and rollback.
Why it matters	Production failures can invalidate the privacy story even when the PET is technically sound.
Current workaround	Borrow generic security launch checklists.
Why the workaround is insufficient	Generic checklists rarely ask about privacy units, output review, collusion assumptions, attestation, or DP accounting.
What good progress would look like	A PET-specific production readiness review with controls mapped to PET families.
Difficulty	Medium
Good for	Privacy engineer, systems builder, security engineer
Related PETs	FL, DP, MPC, HE, TEEs, synthetic data
Possible first contribution	Write a readiness checklist for one private inference deployment and include rollback triggers.

Total Cost Of PET Ownership

Field	Card
Problem	How can teams estimate total cost before choosing a PET?
The itch	PET decisions often compare privacy properties but omit engineering effort, participant support, latency, cloud spend, audits, and user education.
Why it matters	Good ideas stall when cost appears after architecture commitment.
Current workaround	Prototype one candidate and extrapolate.
Why the workaround is insufficient	It misses rejected alternatives and operational costs.
What good progress would look like	A cost worksheet that covers build, operate, monitor, audit, scale, and explainability costs for each PET option.
Difficulty	Good first research problem
Good for	Systems builder, benchmark maintainer, privacy engineer
Related PETs	All PETs
Possible first contribution	Compare total cost drivers for HE inference, TEE inference, and client-side inference for one model provider scenario.

When Governance Beats Cryptography

Field	Card
Problem	How can teams tell when the real problem is governance rather than needing a PET?
The itch	Organizations sometimes choose FL, MPC, or clean rooms to avoid hard conversations about purpose, authority, liability, and acceptable use.
Why it matters	A PET can add complexity without solving the actual blocker.
Current workaround	Let legal and technical teams negotiate separately.
Why the workaround is insufficient	The final design may satisfy neither privacy goals nor business needs.
What good progress would look like	A decision checklist that separates data-movement constraints from governance, incentive, and accountability constraints.
Difficulty	Good first research problem
Good for	Policy researcher, privacy engineer, product/security leader
Related PETs	FL, MPC, clean rooms, TEEs
Possible first contribution	Analyze three proposed PET use cases and identify which constraints are technical versus governance-driven.