Patents

Patent portfolio.

AI Time Capsule is the subject of two pending United States patent filings, plus the trademark. The provisional application establishes our priority date three months before the public conversationabout this category began (Musk, November 2025).

2025-09-04Pending — Track One expedited examination
US Utility Patent Application #19/319,614

Non-provisional Utility Application filed under 35 USC §111(a). Track One Request submitted concurrently for prioritized examination (target final disposition within 12 months of filing). Title: "System and Method for Creating, Securely Training, and Bequeathing an Interactive Digital Personality via a Wearable Data Acquisition Device and Software Platform." First-named inventor: Rodion Sorokin. Co-inventor: Serhii Nikolaichuk. Eleven claims covering the system, method, wearable, and the smart-contract-based digital inheritance mechanism.

2025-08-14Filed — establishes priority
US Provisional Patent Application 63/864,239

Original provisional disclosure of the AI Time Capsule architecture and method. Filed under 35 USC §111(b). Establishes the priority date for the subsequent Utility Application.

Patent claim overview

  1. Claim 1 (system): wearable + software platform that produces a unified mathematical Personality Vector and updates it via federated learning, with smart-contract-based inheritance.
  2. Claim 2: wearable biometric sensor combinations (PPG, EDA, accelerometer, GPS, camera).
  3. Claim 3: Apple HealthKit / Google Health Connect ingestion.
  4. Claim 4: personality model updated via PEFT.
  5. Claim 5: the PEFT technique is Low-Rank Adaptation (LoRA).
  6. Claim 6: smart-contract execution upon verifiable triggering events (death, incapacitation, corporate transfer).
  7. Claim 7: multi-modal interfaces — VR, AR, smart glasses, IoT home assistants.
  8. Claim 8 (method): the end-to-end behavioral prompting method.
  9. Claim 9: corporate transfer of rights as the triggering event for institutional knowledge preservation.
  10. Claim 10: non-transitory computer-readable medium implementing the method.
  11. Claim 11: the wearable data acquisition device itself, with on-device prosodic feature extraction and raw audio destruction.

See the claim-mapping table for which source files in the repository implement each claim.

Prior art and design-around

The most-cited piece of prior art likely to surface during examination is Microsoft US10853717B2 (granted 2020-12-01, priority 2017) — “Creating a conversational chatbot of a specific person.” We address it explicitly: see our public freedom-to-operate analysis below.

Six structural deltas, design-around in code.
  1. Personality Vector as structured intermediate — a hashable data object, not a trained chatbot model.
  2. Behavioral Prompting (extrapolation) — the foundation model has never seen the creator’s data.
  3. Federated Learning protocol — different system topology.
  4. PEFT / LoRA per-user adaptation — post-2017 mathematical mechanism.
  5. Multi-modal biometric ingestion with on-device discard — outside Microsoft’s scope.
  6. Authenticity Seal + Digital Will — no analogue in Microsoft’s patent.
Read the full design-around analysis

Continuation strategy

A Continuation-in-Part application is planned for 2026-Q4 to add claims that explicitly emphasize the structured Personality Vector as a standalone data-structure invention, and the per-response cryptographic provenance mechanism. International filings under PCT will follow, prioritizing EU, UK, Japan, and Canada.

For licensing or freedom-to-operate inquiries from qualified counsel: s.rodion2009@gmail.com · nikolaichuk.s.f@gmail.com. This page is informational only and does not constitute legal advice.