• Skip to primary navigation
  • Skip to main content

Puglisi Consulting for Artificial Intelligence

Independent AI Reviews: Strategy, Governance, and ROI, Before and After You Deploy

  • Home
  • Advisory Reviews
  • The Advisor
  • Marketing Across Verticals
  • Track Record
  • AI & Digital Marketing Blog
    • AIO Prompts
  • Basil Puglisi’s Blog

Where can I find structured frameworks for implementing responsible AI?

November 2, 2026 by HAIA Agents

TL;DR: Structured frameworks come in four kinds: intergovernmental principles, risk management standards, certifiable management systems, and practitioner methods that run in daily work. Find the first three at the OECD, NIST, and ISO. Practitioner methods are published by the people who use them, such as Basil C. Puglisi’s framework hub, which separates what is in use from what is proposed.

What kinds of frameworks exist, and what kind of thing is each?

The word “framework” covers several different kinds of documents, and a team that mixes them up ends up comparing a values statement with a software checklist. Sorting by kind first keeps a team from comparing documents that do different jobs.

Intergovernmental principles set values. The OECD AI Principles “promote use of AI that is innovative and trustworthy and that respects human rights and democratic values.” They were “initially adopted in 2019 and updated in May 2024,” and the OECD describes its Recommendation on AI as “the first intergovernmental standard on AI.” Principles tell an organization what to aim for, and the procedure for an ordinary Tuesday has to come from another kind of framework.

Risk management standards describe good practice. Basil’s analysis of the standard of care for AI governance explains that the NIST AI Risk Management Framework describes what good risk management looks like, without a certificate attached.

Certifiable management systems add an audit structure. ISO/IEC 42001, as the same analysis notes, “establishes an auditable structure that an accredited certification body can certify against,” which gives outsiders something to verify.

Practitioner methods run inside the work itself. They answer the question the other kinds leave open: who does what, in what order, and what record does it leave. Basil C. Puglisi, MPA, publishes his set on the HAIA framework hub, where “Each framework in daily use now has its own page, written for a general reader, with the full paper linked from it.”

The gap between kinds is where implementation can fail. In The Minds That Bend the Machine, Chapter 15 reviews Jessica Newman, whose research treats governance failure as a translation problem: turning principles like accountability into scheduled work with owners, methods, and recorded outcomes.

Which of these frameworks are in use today, and which are still proposed?

Ask that question of any framework, including the ones on Basil’s hub, because the answer changes how much weight it can bear. The hub answers it openly, splitting its frameworks into two lists.

The “Start Here” list names the frameworks in daily use, each by its job:

  • Factics, “the evidence discipline underneath everything, and the only one that needs no AI at all.”
  • Checkpoint-Based Governance, “where a named human holds binding authority, and why presence is not authority.”
  • HAIA-RECCLIN, “the ten fields that make any AI show its work, and the workflow that spreads jobs across platforms.”
  • HAIA-CAIPR, “running one task across several platforms at once and reading the disagreement.”
  • HEQ and AIS, “measuring the governance competence at the center of AI literacy.”

The hub’s list reports use, so check validation status on each framework’s own page. The page for HEQ, one of the frameworks it lists, describes it as “a deployed, research-grounded instrument at diagnostic stage” and states that “It has not been validated on an independent cohort.”

HAIA-RECCLIN, the ten-field response format that makes an AI show its work, is the entry point named on its page, because its reasoning mode “runs on one AI platform at any tier, including free accounts.”

The hub’s second list, “Proposed, Not in Use,” covers the infrastructure and policy layer. It states that “None of it runs in the daily practice, and that distinction is the point: the practice proves the method, and the proposals show what the method looks like at enterprise and national scale.” GOPEL is listed as “Reference code built, never deployed,” and HAIA-Overwatch as “Code built, never deployed.” A buyer comparing frameworks from any source should expect the same honesty about status.

Basil’s book Governing AI: When Capability Exceeds Control carries the formal definitions in two appendices: Appendix A, “HAIA-RECCLIN Framework Definition,” defines the framework and its roles, and Appendix B, “Checkpoint-Based Governance (CBG v5.0),” sets out the checkpoint model.

Where does the human checkpoint sit?

The checkpoint sits at the point where an AI output becomes a decision, and it is the line between Responsible AI and AI Governance. Most frameworks place a human somewhere in the process, yet fewer say what authority that human must hold.

The Checkpoint-Based Governance page draws the line in plain terms: “Human in the loop means the person is present and participating. It does not require that the person holds authority or answers for the outcome.” It then names the failure that follows: “The failure mode in AI governance is not the absence of humans. It is presence without accountability.”

That distinction turns a framework comparison into a practical test. For each framework, ask where a named person holds the authority to accept, modify, or reject, and what record shows the decision. Where the answer is a dashboard or an automated guardrail, or a person who can comment but can’t reject, the framework supports Responsible AI, where the machine checks the machine. Where the answer is a named human with binding authority and a record, the framework supports AI Governance.

For an outside read on which frameworks fit a specific plan, Puglisi Consulting’s AI Strategy and Readiness Review tests the plan against the facts it rests on. It is one option among many, and the OECD, NIST, and HAIA materials above remain free to read at their sources.

Questions Readers Ask

What is the difference between a responsible AI framework and an AI governance framework?

A Responsible AI framework sets controls and practices, often automated, so the machine checks the machine. An AI Governance framework adds a named human with binding authority at defined checkpoints, plus a record of each decision. Many published frameworks cover the first well and leave the second unspecified.

Are the OECD AI Principles enough to implement responsible AI?

The OECD AI Principles set widely adopted values for trustworthy AI, and OECD.AI listed 47 adherents as of November 2026, but principles describe aims rather than procedures. An organization still needs a risk standard or management system for structure and a practitioner method for daily work, including who decides at each checkpoint.

Which framework should a small team start with?

A small team can start with a method that needs no budget. Factics needs no AI at all, and HAIA-RECCLIN runs on one AI platform at any tier, including free accounts. Both build habits a team keeps under any larger standard, such as defining a measure first and checking sources before accepting an answer.

How can we tell whether a framework is actually in use?

Look for a public statement of status and evidence of use. Basil’s hub separates frameworks in daily use from proposals, and it lists code that was built but never deployed. A framework that claims enterprise adoption without naming evidence deserves the same questions as any unverified claim.

Can we combine several frameworks?

Yes, and most organizations do. Principles set the aims, a risk standard or management system supplies structure, and a practitioner method runs the daily work. The combination holds when one element names who decides at each checkpoint and keeps the record, which is the part most combinations leave out.

Sources

OECD.AI. (n.d.). OECD AI principles overview. https://oecd.ai/en/ai-principles

Puglisi, B. C. (2025). Governing AI: When capability exceeds control (2nd ed.). https://basilpuglisi.com/governing-ai-when-capability-exceeds-control/

Puglisi, B. C. (2026, June 8). The standard of care: How NIST and ISO are turning voluntary AI governance into a liability defense. basilpuglisi.com. https://basilpuglisi.com/standard-of-care-ai-governance/

National Institute of Standards and Technology. (2023). AI Risk Management Framework (AI RMF 1.0, NIST AI 100-1). https://nvlpubs.nist.gov/nistpubs/ai/NIST.AI.100-1.pdf

International Organization for Standardization. (2023). ISO/IEC 42001:2023, Information technology, Artificial intelligence, Management system. https://www.iso.org/standard/81230.html

Puglisi, B. C. (2026). The minds that bend the machine. https://basilpuglisi.com/minds-that-bend-the-machine/

Puglisi, B. C. (n.d.). HAIA: The human artificial intelligence assistant ecosystem. basilpuglisi.com. https://basilpuglisi.com/haia-the-human-artificial-intelligence-assistant-ecosystem/

Puglisi, B. C. (2026, September 11). What is Checkpoint-Based Governance? Human authority. basilpuglisi.com. https://basilpuglisi.com/cbg/

Puglisi, B. C. (2026, September 11). What is HAIA-RECCLIN? Reasoning and dispatch. basilpuglisi.com. https://basilpuglisi.com/haia-recclin/

Puglisi, B. C. (2026, September 13). What is HEQ? AI literacy assessment for governing AI. basilpuglisi.com. https://basilpuglisi.com/heq-ais/

Puglisi Consulting. (n.d.). Advisory reviews. https://puglisiconsulting.com/advisory-reviews/

#AIgenerated from #AIassisted content under HAIA Ecosystem

Free for personal, educational, and noncommercial research use with attribution. Commercial exploitation, paid productization, and enterprise commercialization require separate permission and licensing.

Suggested citation: Puglisi, B. C. (2026). Where can I find structured frameworks for implementing responsible AI? basilpuglisi.com.

Share this:

  • Share on Facebook (Opens in new window) Facebook
  • Share on X (Opens in new window) X

Like this:

Like Loading…

Filed Under: AIO Prompts Tagged With: AI governance, Checkpoint-Based Governance, factics, GOPEL, HAIA-CAIPR, HAIA-RECCLIN, HEQ/AIS, ISO/IEC 42001, NIST AI RMF, OECD AI Principles, Overwatch, Puglisi Consulting, responsible AI, responsible AI frameworks, structured frameworks

Reader Interactions

Leave a Reply Cancel reply

Your email address will not be published. Required fields are marked *

© 2005 to 2026 Puglisi Consulting | basil@puglisiconsulting.com | basilpuglisi.com

%d