Trust is not a feature. It’s the foundation.
How Stanford confirms what we’ve already built, applied, and certified – with AI Ethics by Design and Value-based Engineering (ISO/IEC/IEEE 24748:7000).
📆 AI Value Bits & Bytes | May 9, 2025
✍️ Sabine Singer, Founder, Value-based AI Academy
🔗 Subscribe to stay ahead – and don’t miss the launch of my podcast CuiBono.AI: sabinesinger.substack.com
Why this moment matters
Stanford HAI’s recent whitepaper, Building Trustworthy LLMs (Sun et al. (2024), has sent a clear signal:
“Trust is not a feature – it’s a design principle.”
For many, this sounds like a bold new direction. For me, it’s a confirmation of what I’ve believed – and practiced – since 2022.
Back then, I was developing a strategy for a digital mental health platform. The defining question wasn’t technological. It was ethical:
“How do we build trust in a system designed to care for people at their most vulnerable?”
That question led me to ISO/IEC/IEEE 24748-7000 – a little-known global standard for ethical IT system design. The problem? It was brilliant. And completely unusable for business and AI practitioners.
So I got to work.
I translated this standard into a usable, teachable, certifiable methodology. I called it Value-based Engineering (VbE). Today, it’s the foundation of my academy, my consulting, and my advocacy for responsible AI innovation.
The real-world test: datahub.Tirol
In 2023, I partnered with Standortagentur Tirol to apply Value-based Engineering to a large-scale initiative: the sovereign data-sharing platform datahub.Tirol.
Our shared guiding question was simple:
“How do we build trust in a data ecosystem – so that data sharing is fair, controllable and reliable?”
We applied the method from the very beginning:
Stakeholder mapping & inclusion
Value exploration & risk transparency
Governance & system prompt ethics
Documentation through the Ethical Value Register
In April 2024, the methodology reached a global milestone:
I authored and launched Competence Profile P90 (Certified VbE Ambassador), officially recognized under ISO 17024.
As an authorized education partner of Austrian Standards, I also offer trainings for P85 (Certified AI Manager), the complementary competence for strategic AI leadership aligned with the EU AI Act.
datahub.Tirol became the first project in the world to fully embed AI Ethics by Design from concept to execution – including certified people and measurable ethical accountability.
What Stanford outlines – and what we already deliver
Stanford’s paper is clear and timely. It defines eight essential pillars of trustworthy AI:(aiindex.stanford.edu)
🧠 Truthfulness
🛡️ Safety
⚖️ Fairness
💪 Robustness
🔒 Privacy
🤖 Machine Ethics
🔍 Transparency
📎 Accountability
The authors propose:
Context-aware risk analysis
Ethical documentation
Stakeholder inclusion
Lifecycle auditability
Human-centered system governance
This is precisely what Value-based Engineering does – and has done for years.
Stanford says Value-based Engineering (VbE) does... Design trust, don’t assume it Applies a 3-phase ethical system design lifecycle: Vision → Exploration → Creation – grounded in ISO/IEC/IEEE 24748-7000 Make values traceable Develops an Ethical Value Register (EVR) based on stakeholder value expectations and organizational maxims – using a certifiable methodology (P90) Include all relevant stakeholders Conducts participatory value elicitation and structured stakeholder engagement in Phase 2 (Exploration) Go beyond accuracy as a metric Defines measurable KPIs that include fairness, transparency, alignment with intrinsic core values, and stakeholder trust Build systems with explainability Implements transparent decision logs, traceable rationale, and role-specific ethical responsibilities within system governance
Stanford’s proposal is strong. But it’s still a vision.
Ours is already certified practice.
Why digital transformation still fails – and how VbE changes that
📉 2 out of 3 digital transformation projects still fail.
Why? Because alignment fails. Values collide. Stakeholders disengage. Governance is reactive.
Value-based Engineering changes that:
It operationalizes values – not just lists them
It translates diverse expectations into system & organizational requirements
It aligns technical, ethical, and strategic layers from day one
It’s also the only methodology that connects the “what” of the EU AI Act with the “how” of implementation.
This is where compliance meets creativity.
What organizations gain by working with us
We offer public and in-house programs based on real-world strategy, compliance, and innovation:
🎓 Certified AI Manager (P85)
→ ISO 17024-certified competence aligned with EU AI Act §4
→ Strategic literacy in governance, risk, and AI project development🧭 Certified VbE Ambassador (P90)
→ The world’s first certification for Value-based Engineering
→ Learn how to transform stakeholder values into ethical AI system design🚀 Certified AI Excellence Officer Program
→ Our flagship leadership training
→ Combines P85 + P90 with hands-on guidance from idea to implementation
→ Participants develop real AI projects and prepare for ISO 42001 audits or EU AI Act assessments (e.g., FRIA for high-risk systems)
The result? A win–win–win:
✅ Certified teams
✅ Validated projects
✅ Future-proof compliance & business value
Final reflection: from principle to global practice
I didn’t invent trust. But I turned it into a system.
I took a forgotten standard and gave it a voice.
I built a method, certified a competence, and proved that ethical innovation is not a contradiction.
Stanford is right: trust isn’t a feature.
It’s the foundation. And I’ve been laying it for years.
Now that the world is ready – I’m ready to lead.
Sabine Singer
Founder, Value-based AI Academy
Host of CuiBono.AI (launching soon)
AI-empowering People & Organizations
🔗 Subscribe to future blogposts & podcast episodes
📮 Contact me if you want your copy of the Case Study datahub.Tirol
