Skip to content
Talk to AI StandardsTalk to AI Standardsopens in a new tab
Luxury AI Standards atrium at dawn: floor-to-ceiling windows on the left open onto misty mountains and a lake, sofas sit on polished marble at bottom left, a curved white column and glass tower hold a glowing glass helix sculpture at centre, and the cyan circuit AI STANDARDS logo glows on the dark stone wall at right, with warm LED cove lines along the ceiling curves and the planter edge.

Essays & dispatchesOn how intelligent systems should behave

AI Standards

Writing on uncertainty, accountability and sovereignty in AI, and who answers when a system gets it wrong.

What we examine

Core Questions

  • Accountability

    When an AI system acts for an organisation, who owns the outcome, and how that ownership is written down before anything goes wrong.

  • Uncertainty

    Architecture that makes what a model does not know visible and enforceable, instead of hiding it behind confident prose.

  • Sovereignty

    Running intelligence on your own terms: the operational decisions and tradeoffs behind a system you actually control.

  • Agents

    AI that acts, not just answers. What changes when software can send, buy, book and decide on your behalf.

  • Records & Truth

    Immutability establishes a record. It does not establish truth. Provenance, audit and the limits of both.

  • Human Judgment

    What we keep, and what we risk losing, when machines start doing more of the thinking.

22
Published essays

The arc of the argument

How the Essays Build

  1. Why I Started Building AI Systems Instead of Just Using Them
    Feb 2026

    Starting point

    Why I Started Building AI Systems Instead of Just Using Them

    Using AI is easy. Making it dependable in a real workflow is another discipline.

  2. The Problem With AI That Always Says Yes
    Feb 2026

    The failure mode

    The Problem With AI That Always Says Yes

    An assistant that refuses to acknowledge limits is not helpful. It is dangerous to depend on.

  3. A Blockchain Can Preserve a Wrong Answer Forever
    Feb 2026

    Records are not truth

    A Blockchain Can Preserve a Wrong Answer Forever

    Immutability establishes a record. It does not establish truth.

  4. The Architecture of an AI That Knows What It Doesn't Know
    Mar 2026

    The architecture

    The Architecture of an AI That Knows What It Doesn't Know

    Good AI architecture makes uncertainty visible and enforceable.

  5. Building Sovereign AI: The Distance Between a Vision and a Working Product
    Mar 2026

    Sovereignty

    Building Sovereign AI: The Distance Between a Vision and a Working Product

    Sovereignty is not branding. It is a set of operational decisions and tradeoffs.

Essays

  1. SecurityMicrosoft Purview & Multi-Cloud AI Runtime GovernanceMicrosoft Purview governs M365 data access, not multi-cloud AI inference. Learn how CISOs add zero-egress runtime proof across AWS, Azure, and OpenAI.Joe Quenneville5 min
  2. AI GovernanceYour AI Passed the Pilot. Who Controls What Happens Next?An experiment in Pong makes an enterprise governance problem visible: an agent can perform well under one set of conditions and keep its permission to act after those conditions change.Anwesh Rath6 min
  3. StandardsThe FAA Ground Stop & Critical Infrastructure AI Proof72 hours after a nationwide FAA AI rollout, overloaded routers grounded U.S. flights for 5 hours. Here is why critical infrastructure needs AI proof.Joe Quenneville9 min
  4. StandardsDigit 5 & the $300M Humanoid Safety Certification GapAgility Robotics' $300M Digit 5 backlog hinges on 2027 safety proof. Here is how AIS turns conditional Physical AI orders into bankable revenue.Joe Quenneville7 min
  5. Sovereign AIAnthropic's S-1 & Outcome Pricing: A CEO Warning on AIAnthropic's 318-page S-1 and Dario Amodei's pay-for-results remarks prove cloud AI labs want equity in your IP. Here is the CEO sovereign alternative.Joe Quenneville8 min
  6. GRCSEC FY2027 AI Exams & Treasury FS AI RMF: 9 Fiduciary TestsHow the SEC FY2027 Examination Priorities and U.S. Treasury FS AI RMF reshape fiduciary AI compliance, and how AI Standards answers all 9 examiner tests.Fran Horvath8 min

Leadership

Veteran professionals. Real experience.

No junior analysts running your deployment. The people who designed the system are the people who build yours.

  • Anwesh Rath

    Co-Founder · CTO & Chief Scientist

    Architects and builds the platform end to end: sovereign model stacks, governed inference, and the cryptographic governance layer that lets organizations prove exactly what their AI did. Leads the research turning “trust us” compliance into receipts you can independently verify.

  • Joe Quenneville

    Co-Founder & CEO

    Joe Quenneville is the Founder and CEO of AI Standards, bringing over 35 years of technology leadership to a single conviction: nearly every AI failure, risk, and low-quality result traces back to the absence of standards. He leads the company from decades of leadership across enterprise technology, cybersecurity, and compliance.

  • Fran Horvath

    Co-Founder & CVO / Chief Vision Officer

    Fran Horvath identified the AI governance gap before the market had language for it, and designed the systems to close it. She architects the strategic, operational, and financial frameworks behind AI Standards’ deployment model.

Write

New essays and dispatches go out on the feed. Write if a piece should exist and does not.

Talk to AI StandardsTalk to AI Standards
Talk to AI StandardsTalk to AI Standardsopens in a new tab