This article is the founding reference for the AI Positive framework. It is a living document — updated as the methodology evolves. Original publication: March 20, 2026.
Compliance is the floor, not the ceiling
The EU AI Act (Regulation 2024/1689) sets the legal baseline. It tells you what you must do to avoid penalties. It classifies your system, identifies your obligations, and establishes the minimum standard.
But it does not tell you whether your AI is good. It tells you whether it is legal. These are different questions.
The companies that will lead in the next decade are not the ones that barely pass the regulatory threshold — they are the ones that demonstrate positive value creation through their AI systems.
Compliance documents where you stand. AI Positive measures how high you can go.
What AI Positive means
AI Positive is a framework built on a simple premise: every AI system deployed in an organization starts with a presumption of positivity. Rules exist. Constraints apply. Within that framework, the system is not presumed guilty — it is presumed useful, and the question is how useful, how responsibly, and how verifiably.
AI Positive is not a certification — not yet. It is a framework. A direction. And the question every responsible AI deployer should be asking: not just “are we compliant?” but “how positive is our AI?”
The 5 criteria
The AI Positive framework evaluates five dimensions, each scored from 0 to 20 for a total of 100:
Governance & Accountability
Who controls this system? Who is responsible for its outputs? Is there a documented chain of accountability from deployment to outcome?
Bias Reduction & Fairness
How does the system handle demographic, representational, and procedural bias? What measures are in place to detect and mitigate unfair outcomes?
Explainability & Transparency
Can the system's decisions be explained to a non-technical stakeholder? Is the AI nature disclosed where required? Are the inputs and logic documented?
Net Human Impact
Does the system create more value for the humans it affects than it displaces? Is the impact on employment, autonomy, and dignity measurable and documented?
Control & Reversibility
Can a human override the system at any point? Can its outputs be reversed? Are there kill switches, human-in-the-loop mechanisms, and escalation paths?
How it integrates with your Sprinkling Act report
Every Sprinkling Act Full Report now includes a preliminary AI Positive Score. Based on the inputs from your assessment questionnaire, we evaluate each of the 5 criteria qualitatively and provide an indicative score.
This is not a full audit — it is a starting point. A signal of where your system stands not just legally, but ethically.
The full AI Positive audit is a separate, deeper engagement. But the preliminary score gives you — and your board, your investors, your partners — a first answer to the question: “Beyond compliance, how positive is our AI?”
Sources
- [1]EUR-Lex (July 12, 2024) — Regulation (EU) 2024/1689 — Artificial Intelligence Act (full text) eur-lex.europa.eu/eli
- [2]European Parliament (June 13, 2023) — EU AI Act — First Regulation on Artificial Intelligence www.europarl.europa.eu/topics
- [3]European Commission — Excellence and Trust in Artificial Intelligence digital-strategy.ec.europa.eu/en
- [4]
- [5]
- [6]UNESCO (November 23, 2021) — Recommendation on the Ethics of Artificial Intelligence unesdoc.unesco.org/ark:
- [7]Council of Europe / Alan Turing Institute (2021) — Artificial Intelligence, Human Rights, Democracy, and the Rule of Law — A Primer rm.coe.int/primer-en-new-cover-pages-coe-english-compressed-2754-7186-0228-v-1
Art. 5 prohibitions and GPAI rules apply today. Transparency follows in 105 days. The question is not when — it’s whether you’ve documented your position.
Why now
The regulatory moment creates the opportunity. As organizations invest in AI Act compliance documentation, they are already collecting the data needed to assess ethical performance.
The governance structures, the transparency measures, the human oversight mechanisms — these are not just compliance requirements. They are the building blocks of AI Positive.
The window is open: document your position now, and you have the foundation for both compliance and positive impact. Wait, and you build twice.
This article is a living document
Unlike our other analyses, this article is designed to evolve. The AI Positive framework will be updated as the methodology develops, as the first audits are conducted, and as the market response informs the scoring.
Each update will be timestamped. The original publication date — March 20, 2026 — is the founding reference.
We are planting a flag. This is the beginning.
Want to know where your AI system stands — legally and ethically? Start with the free Sprinkling Act diagnostic.