arXiv:2605.29524v1 Announce Type: cross Abstract: Relay and reseller APIs increasingly intermediate access to large language models (LLMs), but users have no direct way to verify that a claimed endpoint is actually serving the advertised model. We introduce KBF, a low-cost black-box auditing protocol that fingerprints model APIs using stable numerical recall near the knowledge boundary. Across 16 production LLM endpoints, KBF flags all 155 economically relevant substitutions without rejecting any same-model controls, remains stable under deployment variation, detects high-separation mixed-rout
The proliferation of LLM APIs and the increasing reliance on intermediaries necessitate robust methods for verifying model authenticity and preventing fraudulent substitutions.
A strategic reader should care because this development addresses a critical trust and security issue in the AI supply chain, impacting enterprise adoption and regulatory oversight.
Users can now independently verify the specific LLM being served by an API endpoint, reducing the risk of fraud and improving transparency in outsourced AI services.
- · LLM Customers
- · Independent AI Auditors
- · Ethical AI Developers
- · Open-source LLMs
- · Fraudulent API Resellers
- · Undisclosed Model Substitutions
- · Black-box API Providers
Increased trust in LLM API services due to verified authenticity.
Potential for new regulatory requirements for API transparency and model verification.
Drives greater adoption of LLMs in critical applications where provenance and integrity are paramount.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.AI