For MSF Working Group

These are the prioritized recommendations for MSF AI x Metaverse Working Group action.

Priority 1
MEDIUM Effort Q2 2026

Propose MSF_ai_generation_metadata glTF extension

No existing work addresses AI generation metadata. This is a greenfield opportunity with high impact on reproducibility and interoperability.

Scope:

  • Generation prompt/parameters
  • Seed and random state
  • Model version and configuration
  • Quality metrics
  • Generation timestamp

Path: Vendor extension -> KHR ratification

Priority 2
MEDIUM Effort Q3 2026

Define geometry quality validation profiles

Addresses topology errors that block production use. Enables certification of AI-generated content.

Scope:

  • Non-manifold detection
  • Physics compatibility checks
  • UV quality assessment
  • LOD requirements
  • Polygon budget compliance
Priority 3
LOW Effort Q1 2026

Formally engage C2PA for 3D workflow support

Critical gap in provenance tracking. MSF can advocate on behalf of the industry.

Scope:

  • Establish MSF liaison with C2PA
  • Document 3D workflow requirements
  • Propose glTF/USD embedding mechanisms
Priority 4
LOW Effort Ongoing

Support KHR_gaussian_splatting ratification

Enable neural representation interchange. Monitor and contribute to Khronos work.

Scope:

  • Track PR #2490 progress
  • Contribute use cases from AI generation
  • Support compression work (SPZ_2)
Priority 5
HIGH Effort Q4 2026

Convene world package format discussion

Complex multi-SDO coordination needed. MSF is uniquely positioned to facilitate.

Scope:

  • Bring together Khronos, AOUSD stakeholders
  • Define requirements for portable world packaging
  • Identify lead SDO and scope

Priority Summary Table

PriorityRecommendationRationaleEffortTimeline
1Propose MSF_ai_generation_metadata glTF extensionNo existing work; greenfield opportunity; high impactMEDIUMQ2 2026
2Define geometry quality validation profilesAddresses topology errors blocking production useMEDIUMQ3 2026
3Formally engage C2PA for 3D workflow supportCritical gap; MSF can advocate on behalf of industryLOWQ1 2026
4Support KHR_gaussian_splatting ratificationEnable neural representation interchangeLOWOngoing
5Convene world package format discussionComplex multi-SDO coordination; MSF uniquely positionedHIGHQ4 2026

Coordination Opportunities

These are opportunities for MSF to coordinate between SDOs and industry.

OpportunitySDOs InvolvedMSF RoleExpected Outcome
AI Generation Metadata ExtensionKhronos, MSFConvene, propose requirementsVendor extension -> KHR path
C2PA 3D WorkflowC2PA, KhronosFacilitate, advocate3D embedding specification
Neural-to-Mesh ConversionKhronos, AOUSDObserve, contribute use casesQuality standards for conversion
Multiplayer DeterminismMSF (new)LeadRequirements specification

AI Generation Metadata Extension

Current State: No work exists MSF Role: Lead proposal

Steps:

  1. Draft requirements document (Q1 2026)
  2. Propose vendor extension to Khronos (Q2 2026)
  3. Gather implementer feedback (Q2-Q3 2026)
  4. Path to KHR ratification (2027+)

C2PA 3D Workflow

Current State: No 3D support in C2PA MSF Role: Advocate

Steps:

  1. Establish liaison (Q1 2026)
  2. Document requirements
  3. Propose embedding mechanisms
  4. Support specification work

Multiplayer Determinism

Current State: No specification exists MSF Role: New work item

Steps:

  1. Define problem scope
  2. Research technical feasibility
  3. Draft requirements
  4. Identify implementation partners

Research Priorities

These research questions should be addressed to inform standards work.

PriorityResearch QuestionMethodResources Needed
1What metadata is essential for AI generation reproducibility?Creator surveys, workflow analysis3-month study, industry partners
2What topology quality thresholds are acceptable for physics?Benchmark with game studiosTechnical study, engine expertise
3Can deterministic generation be achieved across GPU vendors?Technical feasibility studyNVIDIA, AMD, Intel engagement

Research Question 1: Essential Metadata

Why it matters: Informs MSF_ai_generation_metadata schema design

Approach:

  • Survey creators using AI generation tools
  • Analyze existing metadata exports from tools
  • Identify minimum viable metadata set
  • Validate with reproducibility testing

Research Question 2: Topology Quality Thresholds

Why it matters: Defines quality validation profiles

Approach:

  • Partner with game studios
  • Test AI-generated meshes in production engines
  • Measure physics failure rates
  • Define acceptable thresholds

Research Question 3: Deterministic Generation

Why it matters: Enables multiplayer AI-generated content

Approach:

  • Technical review of AI model architectures
  • GPU vendor consultation
  • Test reproducibility across hardware
  • Document achievable determinism levels

Use Case Development

These use cases should be prioritized for standards work.

Use CasePriorityReadinessDependencies
Batch AI World Generation for GamesHIGHReadyQuality validation profiles
Runtime AI Environment GenerationMEDIUMNeeds ResearchDeterminism specification
AI Generation Metadata InterchangeHIGHReadyExtension proposal
Neural-to-Mesh Conversion WorkflowsHIGHBlockedKHR_gaussian_splatting ratification

Batch AI World Generation

Status: Ready for standards work Blocking: Quality validation profiles

Use case: Game developers generate world assets via AI tools, then import into engines for production use.

Runtime AI Generation

Status: Needs research Blocking: Determinism specification

Use case: Games generate content at runtime using AI, requiring reproducibility for multiplayer synchronization.

AI Metadata Interchange

Status: Ready for standards work Blocking: Extension proposal

Use case: AI-generated assets carry metadata about generation parameters, enabling reproducibility and provenance.

Neural-to-Mesh Conversion

Status: Blocked Blocking: KHR_gaussian_splatting ratification

Use case: Content created in neural-native tools (Gaussian splats) converted to mesh for engine use.


Roadmap Summary

Q1 2026

  • Establish C2PA liaison
  • Begin metadata research study

Q2 2026

  • Propose MSF_ai_generation_metadata extension
  • Complete metadata research

Q3 2026

  • Define quality validation profiles
  • Begin topology threshold study

Q4 2026

  • Convene world package format discussion
  • Complete quality research

2027+

  • Path KHR ratification for AI metadata
  • Address world package format
  • Support C2PA 3D specification