LinkedIn·Monday, 3 August 2026·23d ago

A digital twin is only as trustworthy as the data, systems, processes and people that make it work. We built a rulebook for trust, before…

3DxVERSE Project

380 followers

A digital twin is only as trustworthy as the data, systems, processes and people that make it work. We built a rulebook for trust, before the pilots even started. Most digital projects bolt on privacy and ethics at the end. We didn't. #3DxVERSE built a Trust Framework from day one. It's the rulebook for how our digital twins can be trusted: who is trusted, for what purpose, under what conditions, and with what assurance. Here's what that means in practice: → 𝐃𝐚𝐭𝐚 𝐢𝐬 𝐠𝐨𝐯𝐞𝐫𝐧𝐞𝐝, 𝐧𝐨𝐭 𝐣𝐮𝐬𝐭 𝐜𝐨𝐥𝐥𝐞𝐜𝐭𝐞𝐝. Wherever personal data is involved, there is a declared purpose and a documented lawful basis, and that processing is assessed and improved through a Data Protection Impact Assessment (DPIA). → 𝐒𝐞𝐜𝐮𝐫𝐢𝐭𝐲 𝐢𝐬 𝐛𝐮𝐢𝐥𝐭 𝐢𝐧, 𝐧𝐨𝐭 𝐛𝐨𝐥𝐭𝐞𝐝 𝐨𝐧. Weaknesses are caught as the software is built, every component is inventoried, and system logs are tamper-evident and kept for at least 12 months. → 𝐓𝐡𝐞 𝐫𝐮𝐥𝐞𝐬 𝐚𝐫𝐞 𝐰𝐫𝐢𝐭𝐭𝐞𝐧 𝐝𝐨𝐰𝐧 𝐚𝐧𝐝 𝐚𝐮𝐝𝐢𝐭𝐚𝐛𝐥𝐞. Who may do what, for which purpose and to what level of assurance is set out in controlled governance documents so that decisions can be traced and reviewed, not left to discretion. → 𝐏𝐞𝐨𝐩𝐥𝐞 𝐡𝐚𝐯𝐞 𝐝𝐞𝐟𝐢𝐧𝐞𝐝 𝐫𝐨𝐥𝐞𝐬 𝐚𝐧𝐝 𝐜𝐥𝐞𝐚𝐫 𝐚𝐜𝐜𝐨𝐮𝐧𝐭𝐚𝐛𝐢𝐥𝐢𝐭𝐲. Every participant, whether they are a citizen, an organisation, or an operator is identified to a level of assurance that fits the task and holds a defined role, so it is always clear who is responsible for each action. → 𝐄𝐯𝐞𝐫𝐲 𝐀𝐈 𝐭𝐨𝐨𝐥 𝐢𝐬 𝐥𝐨𝐠𝐠𝐞𝐝 𝐚𝐧𝐝 𝐫𝐢𝐬𝐤-𝐫𝐚𝐭𝐞𝐝. Our 11 AI systems, from emissions analytics to energy-label inference, are classified under the EU AI Act's four risk tiers before they're used. → 𝐀 𝐫𝐢𝐠𝐡𝐭𝐬 𝐜𝐡𝐞𝐜𝐤 𝐛𝐞𝐟𝐨𝐫𝐞 𝐝𝐞𝐩𝐥𝐨𝐲𝐦𝐞𝐧𝐭. Higher-risk systems must pass a Fundamental Rights Impact Assessment (FRIA). A hard gate that cannot be waived. → 𝐂𝐢𝐭𝐢𝐳𝐞𝐧𝐬 𝐚𝐫𝐞 𝐭𝐨𝐥𝐝 𝐰𝐡𝐞𝐧 𝐀𝐈 𝐢𝐬 𝐢𝐧𝐯𝐨𝐥𝐯𝐞𝐝. From August 2026, that's a legal requirement under AI Act Article 50. We're already there. → 𝐅𝐚𝐢𝐫𝐧𝐞𝐬𝐬 𝐢𝐬 𝐭𝐞𝐬𝐭𝐞𝐝. Algorithmic fairness testing is mandatory for citizen-facing AI. Results will be published in a public AI register. → 𝐁𝐮𝐢𝐥𝐭 𝐭𝐨 𝐛𝐞 𝐫𝐞𝐮𝐬𝐞𝐝. The framework is portable into the LDT Toolbox and the CitiVERSE EDIC. Other cities can adopt it without rewriting the rules. The framework extends iSHARE Foundation's Trust Framework for data rights, adding capabilities for digital twins: autonomous machines, AI agents, real-world devices like energy meters, and a citizen engagement layer. It sits within the EU's architecture for governing data spaces (DS4SSCC, GAIA-X, IDS-RAM) and digital identity (eIDAS 2.0 and digital wallets). Digital public infrastructure has to be dependable by default. People can't simply choose a different provider. The burden of trust sits with us. What's one thing you'd want to know before trusting a digital twin with your data?

View on LinkedIn