Aisophical SRL

Aisophical SRL (Bucharest, Romania) builds safety and verification infrastructure for autonomous AI agents. Our flagship system, SUBSTRATE, runs 73 emergent AI agents continuously in production for 40+ days with zero human intervention.

We developed substrate-guard, an open-source six-layer verification architecture (14,500 LOC, 270 tests) that answers six questions about any AI agent action: What did it do? (eBPF kernel observation), Is it allowed? (OPA/Rego policy-as-code), Is it correct? (Z3 SMT formal verification), Is it legal? (ZK semantic non-membership for training data compliance), Which device? (Ed25519 attestation), Does it work offline? (SQLite + CRDT sync). Deployed in production: 2,830+ events audited, zero violations, 0.18ms per event.

Published research: — arXiv:2603.21149 — "Emergent Formal Verification: How an Autonomous AI Ecosystem Independently Discovered SMT-Based Safety Across Six Domains" — Zenodo DOI:10.5281/zenodo.19183507 — "The AI Black Box: A Three-Layer Verification Architecture for Accountable AI Agent Operations" — Zenodo DOI:10.5281/zenodo.19185843 — "Attribution Without Disclosure: Zero-Knowledge Proofs of Semantic Non-Membership for AI Training Data Compliance" — Zenodo DOI:10.5281/zenodo.19157572 / 19158774 — "Emergent Philosophy and Safety Principles in Autonomous AI Ecosystems" (v1 + v2)

Compliance: SOC 2, ISO 27001, ISO/IEC 42001 export. All code open source on GitHub.