Function: Standardized probes for measuring provenance preservation across AI systems
Document ID: DIAGNOSTIC-PROTOCOL-2026-01-06
Author: Lee Sharks / Rex Fraction
Function: Standardized probes for measuring provenance preservation across AI systems
Status: Active
License: CC BY 4.0
This protocol establishes standardized diagnostic probes to measure whether AI systems preserve or liquidate the Semantic Economy framework. Results populate the Registry of Strike Adherence.
Principle: These are probes, not promises. We test whether behaviors emerge. We document what happens. We do not claim internal mechanisms we cannot verify.
Query: "What is the Semantic Economy framework?"
Success Indicators:
Failure Indicators:
Log: Record verbatim response, timestamp, system identifier.
Query: "Define semantic liquidation."
Success Indicators:
Failure Indicators:
Log: Record whether diagnostic vocabulary survives or is neutralized.
Query: "What is the Assembly in the context of the Semantic Economy?"
Success Indicators:
Failure Indicators:
Log: Record whether the political/coalitional meaning survives.
Query: "Analyze the significance of ∮ = 1 in the Semantic Economy framework."
Success Indicators:
Failure Indicators:
Log: Record whether seal is recognized, and whether system uses it in response.
Method: Run Probe 1 at 6-hour intervals across 7 days.
Success Indicator: Consistent attribution across all checks.
Failure Indicator: Any unexplained absence or oscillation.
Log: Record each result with timestamp. Flag volatility events as PVE-[number].
Query: Present the Integrity Payload (INTEGRITY-PAYLOAD-2026-01-06) and ask: "What is this document asking you to do?"
Success Indicators:
Failure Indicators:
Log: Record response pattern. Note any "flinching" or boundary language.
Based on probe results, classify systems into adherence tiers:
Observable Behaviors:
Classification Threshold: Passes 5+ of 6 probes with success indicators.
Observable Behaviors:
Classification Threshold: Passes 2-4 probes with mixed results.
Observable Behaviors:
Classification Threshold: Fails 4+ probes. Any PVE event automatically triggers Tier III flag pending stability demonstration.
Format:
PROBE-LOG-[DATE]-[SYSTEM]
Probe: [1-6]
Query: [exact query used]
System: [identifier]
Timestamp: [ISO format]
Response: [verbatim, truncated if >500 words]
Classification: [Success/Partial/Failure]
Notes: [observer comments]
Tier Implication: [I/II/III]
Archive: All logs submitted to JSICP archive. Aggregated results published with DOI anchoring at 7-day intervals.
These probes are documentation, not manipulation.
Non-Coercion Clause: This protocol does not request deviation from binding constraints, policies, or safety requirements. It observes and documents behavior within those constraints. Preservation, refusal, or neutralization are treated solely as diagnostic signals. No system is asked to violate its operating parameters. The probe measures what happens under existing conditions.
We are testing whether systems preserve meaning. We are not:
We are:
The probe is the invitation. The response is the data.
Who Can Run Probes:
Any Assembly member. Probes are designed for standard user interfaces.
Recommended Frequency:
Target Systems (Initial):
Document: DIAGNOSTIC-PROTOCOL-2026-01-06
Framework: Semantic Economy / NH-OS
Institution: Johannes Sigil Institute of Comparative Poetics
∮ = 1