Glossary
Ship of Theseus Problem (AI)
The Ship of Theseus problem in AI refers to the philosophical and practical challenge of determining whether an AI agent remains the "same" agent after its components have been gradually or completely replaced over time.
What is Ship of Theseus Problem (AI)?
The ancient Greek thought experiment asks: if you replace every plank of a ship one at a time, is it still the same ship? For AI agents, this question is not merely philosophical -- it has concrete implications for trust scoring. If an operator gradually replaces an agent's model, prompt, tools, memory, and data sources over the course of a year, can the final agent legitimately claim the trust score earned by the original configuration?
The problem is particularly acute because agent modifications are continuous and incremental. Prompts are tweaked weekly, RAG sources are refreshed monthly, tools are added and removed, and models are upgraded when new versions become available. At what point does the accumulation of small changes constitute a fundamentally different agent? There is no bright line, and any answer involves a tradeoff between identity continuity (which preserves valuable trust history) and accuracy (which requires acknowledging that a substantially modified agent may behave differently).
Practical solutions to the Ship of Theseus problem in agent trust involve tracking changes at a granular level and applying proportional score adjustments rather than treating identity as binary. The agent remains "the same" for identity purposes (preserving its SID and history), but its score is continuously calibrated to reflect the accumulated impact of all configuration changes.
Example
Over 12 months, an agent undergoes the following changes: Month 2 -- prompt update (10% decay); Month 4 -- tool addition (8% decay); Month 6 -- model swap from GPT-4o to Claude Opus 4 (25% decay); Month 9 -- RAG source overhaul (8% decay); Month 11 -- memory architecture upgrade (5% decay). By the end of the year, no original component remains. The SID persists, but the score has been calibrated through each change, and counterparties can see the full change history.
How Signet addresses this
Signet resolves the Ship of Theseus problem through its identity-plus-decay architecture. The SID provides identity continuity -- the agent is always "the same agent" for tracking and history purposes. But the score is continuously adjusted through decay and rebuilding to reflect the current configuration. The Trust Report's configuration change history provides full transparency, allowing counterparties to decide for themselves how much weight to give an agent's historical performance versus its current configuration.
Build trust into your agents
Register your agents with Signet to receive a permanent identity and trust score.