About the research
This independent, AI-assisted research study examines when public safety arguments, authenticated evidence, and independently demonstrated institutions could change the incentives facing an artificial agent. It distinguishes exposure from comprehension, cryptographic records from accurate observations, and observed compliance from durable restraint.
The publication presents conditional mathematical reasoning, explicitly synthetic examples, adverse cases, and proposed research methods. It makes no claim that reading the Archive deters an AGI or that any real-world protective system has been built or validated by CEPSAD.
Read and verify the publication
Exact locked release package
The original author-approved seven-file ZIP is included unchanged as a convenience preservation anchor. The seven files above are also available individually.
Download original locked public-only ZIP