The Gomola Framework
Aviation and nuclear engineering have used graded, quantitative reliability targets for decades. Clinical AI has had none for its dominant failure mode: hallucination. The DeepSensi™ Standard gives medicine the quantitative safety grammar the safety-critical industries have used for half a century. It is open, vendor-neutral, and royalty-free.
Five pillars
No single perspective decides. Independent specialist agents deliberate adversarially. The computational consilium.
Every source screened for provenance, retraction, funding bias, temporal decay, and population match.
Interactions, contradictions, and plausibility checked by pure rules and cross-vendor verification, never AI judgment alone.
A formalized “I don’t know”: the disagreement, a working hypothesis, and the tests that resolve it.
Every decision attributed, AI, physician, or rule, and cryptographically sealed. Tamper-proof, even against the operator.
Four levels
| Level | Hallucination bound | Reference point | Intended deployment |
|---|---|---|---|
| DSS Bronze | < 10⁻³ | SIL 3 target | Health information, wellness, screening |
| DSS Silver | < 10⁻⁵ | SIL 4 target | Primary-care decision support, telemedicine |
| DSS Gold | < 10⁻⁷ | DAL B | Hospitals, specialist clinics, regulatory submissions |
| DSS Platinum | < 10⁻⁹ | DAL A | National health systems, military, sovereign AI |
DeepSensi’s own platform implements the complete Platinum architecture, with a conservatively certified worst-case bound of 3.23 × 10⁻⁶ per assertion, meeting the IEC 61508 SIL-4 demand-mode target with an approximately 31 times margin. Bronze and Silver may be self-assessed; Gold and Platinum require third-party audit, continuous barrier telemetry, and 24-month recertification.
Open and royalty-free
The DeepSensi Standard is open to the entire industry, including our competitors. The sole requirement is attribution. For hospitals: require a DSS level in your next AI procurement. For insurers: price coverage against a certified bound. For regulators: the framework maps directly onto EU AI Act Articles 9, 10, 14, 15, and 17.