Science fiction is becoming infrastructure
What if an AI could change its model, its body, and its world—yet remain recognizably itself?
Persistent Cognitive Identity is our proposal for turning that science-fiction question into a testable systems architecture. We believe we can make it work.
Narrative chapters
PCI identity tests
Buildable conviction
What is Persistent Cognitive Identity?
PCI is a governed continuity layer for AI. It carries the evidence, memory, beliefs, relationships, development rules, embodiment context, and authority needed to judge whether a changing system remains the same operational identity—or has become a branch, a descendant, or a compromised copy.
Three tests that ask harder questions than “Can it seem human?”
Can it seem human?
Does it match this person?
Does its life story make sense?
Is it still the same identity after change?
Three tests to evaluate identity — not just imitation.
The classical Turing Test asks whether a machine can seem human. PCI proposes three new tests — HRFT, SIT, and CCT — that ask whether a system faithfully represents a specific person, owns a coherent biography, and survives change as the same governed identity. Each page compares the test with the classical baseline, shows what evidence it needs, and explains what a result can — and cannot — claim.
HRFTHuman-Referential Fidelity Test
Does this system match this particular person?
The Human-Referential Fidelity Test evaluates how faithfully a consented AI twin represents one specific source human under a declared protocol.
SITSituated Identity Test
Is this agent's behavior uniquely attributable to its actual developmental lineage?
The Situated Identity Test evaluates whether an autonomous agent's behavior is functionally attributable to a specific developmental lineage, enforcing both appropriate knowledge and appropriate ignorance.
CCTCognitive Continuity Test
Is this still the same governed identity after change?
The Cognitive Continuity Test evaluates whether an AI remains the same governed identity after learning, migration, branching, merging, rollback, or a new body.
What this manifesto believes
The claim is ambitious, but it is operational. We can name the parts, build the controls, and design the tests.
Five chapters, one continuous argument.
Science Fiction Is Becoming an Engineering Problem
Why AI identity must become more than a prompt, persona, account, or familiar voice as systems move across models, platforms, and bodies.
What Must Persist When the Model Changes?
Explore the continuity kernel, seven layers of persistent AI identity, and the difference between general model knowledge and lived experience.
Beyond the Turing Test: HRFT, SIT, and CCT
A simple comparison of the classical Turing Test with PCI’s Human-Referential Fidelity Test, Situated Identity Test, and Cognitive Continuity Test.
One Identity, Many Branches, Bodies, and Possible Futures
How an AI identity can migrate, branch, merge, enter new bodies, or become a distinct descendant without silently inheriting authority.
We Believe Persistent Cognitive Identity Can Be Built
A practical PCI roadmap for a continuity kernel, IdentityLineageBench, model migration, adversarial tests, and governed digital succession.
Use this manifesto for the narrative and plain-language test framework. Read the foundational Continuity Kernel paper on arXiv (arXiv:2608.11632), or download the complete technical systems architecture PDF.
Continuity Kernel (arXiv:2608.11632):
PCI Systems Architecture (OpenKedge Research):