Theses on infrastructure,
regulation, and digital truth.
Strategic analysis of AI governance, cryptographic evidence, and European regulatory frameworks, for legal, compliance, and executive management teams.
Section 14 of the AI Act: How to Demonstrate Human Oversight
Section 14 of the AI Act requires that high-risk AI systems be monitored by people who understand them and can shut them down. To demonstrate compliance, a simple “pass” is not enough: it must be possible to trace who reviewed each decision, what they saw, when, and whether they had the authority to reject it.
AI-to-Human Ratio: What It Is and How to Demonstrate It in an Audit
If an EU AI Act auditor, an insurance company, or an ISO 42001 certifier is looking for just one piece of data, it’s the AI-to-human ratio. What it is, how it’s measured, and how to demonstrate it with verifiable evidence.
The first harmonized standard under the AI Act is now available: EN 18286 and the presumption of conformity under Article 17
The European Commission has published the reference to EN 18286:2026—the first harmonized standard under the AI Act—in the Official Journal. Those who implement it are presumed to be in compliance with Article 17 (quality management system) for high-risk systems. It is based on ISO/IEC 42001 and ISO 9001. What changes for CIOs, CDOs, and Legal, and why the presumption is only as valuable as the evidence.
The asset your insurer will require before covering an AI incident
When an AI incident occurs, your insurer won’t ask about your responsible use policy. It will ask if you can prove that the system was operating with active controls in place at the exact moment the damage occurred. What evidence do AI underwriters already require (immutable, third-party verifiable, granular, and continuous), and why doesn’t a GRC system based on policies provide it?
Copyright and AI: Human authorship is no longer presumed. It must be proven.
AI can create, but only a person can be the author. The European Parliament confirms that content generated entirely by AI is not protected by copyright. The challenge now is to prove human contribution.
V-PROOF Wins the 2026 Innovacat Awards in the Osona Jove Category
The jury for the 10th edition of the Premis Innovacat awards has recognized V-PROOF, the verifiable evidence infrastructure for AI governance developed in Vic.
From Policy to Practice: Progress on the Verifiable AI Governance Handbook
Almost all companies that use AI have a policy governing its use. Very few can provide evidence that they are complying with it. On Tuesday, October 20, we published the Verifiable AI Governance Handbook: 22 pages on how to move from policy to proof. This is a step forward.
What Is an AI Registry, and Why Is It the First Step in Any AI Governance Program?
Inventory and AI registry are not the same thing. The inventory lists what you have. The registry manages how what you have behaves—and whether that behavior meets the requirements you must demonstrate. Without a registry, the EU AI Act, ISO 42001, and AI insurance are managed in silos with inconsistent data. It is the infrastructure for everything else.
V-PROOF Expands: Who's in Charge in a World with Millions of AI Models?
Gil Blancafort, co-founder of V-PROOF, participates in Expansión's discussion on the future of AI models. Our perspective: more models require accountability, oversight, and verifiable evidence.
Which version is currently in effect? And who approved it?
In any audit of an ISO management system, two questions always come up: Which revision is currently in effect, and who approved it? All too often, the answer begins with a search for paperwork.
What portion of your code was written by AI? The question is no longer a technical one.
Code assistants on every team, dozens of branches, and a repository that no one knows in its entirety. For technical management, it’s a control issue. For Legal, it’s a matter of authorship and licensing. For the auditor, it’s a matter of evidence.
Which version of your model is in production, and who approved it?
Your data team knows which version is the winning one. What an audit asks is something else: who decided to put it into production, what controls were in place, and exactly which files were running on the day of the incident.
His agent is already taking action. Can he show what he was capable of?
AI agents no longer just respond: they query systems, select tools, and move data. The auditor’s question will be simple: What could that agent do, with what permissions, and who approved it?
The AI Governance Market Is Growing at an Annual Rate of 67.5%: What This Means for Your Company
The AI governance platform market is growing at an annual rate of 67.5%. This is not an optimistic projection—it is the result of four forces that are already at work: the EU AI Act already in effect, publicly reported AI incidents, insurers changing their terms and conditions, and a risk that is dynamic, not static. The window of competitive advantage exists right now.
We are finalists for the 2026 Innovacat Awards
V-PROOF® is one of the 12 finalists in the Osona Jove category of the 2026 Innovacat Awards. On October 6, we will present the jury in Manlleu with a sealed presentation that they can verify themselves using a QR code.
Your code already has a history. Strategic Provenance turns that history into evidence.
Who understands your code, how much of it was written by AI, and what changed in each version. V-PROOF turns your repository into verifiable evidence for CIOs and auditors.
AI Systems Inventory: A Practical Guide to Compliance
Most organizations that launch an AI governance program encounter the same problem: they don’t know exactly which AI systems they’re using. They don’t have a centralized, up-to-date view. Without an inventory, there is no governance—and without governance, there is no compliance with the EU AI Act or ISO 42001.
Why Your GRC Tool Isn't Enough to Govern AI—and What Is
GRC tools document risks, policies, and controls. Governing AI also requires evaluating each system and providing evidence that a third party can verify. This adds a layer of AI governance with verifiable evidence.
The AI seatbelt. And what Volvo understood before anyone else.
In 1959, Nils Bohlin invented something that no one had asked for. Volvo patented it and then gave the patent away—not to make money, but to ensure that trust in the car would become a universal standard. Today, we’re at that same juncture with AI. The problem isn’t technological; it’s one of evidence.
Why the evidentiary infrastructure is not a dat governance —and why that difference matters.
The AI governance market is growing rapidly. And with it comes confusion that has real-world consequences. Every week, new platforms emerge promising “immutable records” and “compliance by design.” They all use the same vocabulary. But not all of them solve the same problem.
