人工智能（AI）政策

Artificial Intelligence (AI) Policy At Arup we see AI, as defined by the OECD, as a pivotal technology that will augment,not replace, our collective and creative intelligence, judgement, and expertise. We aimto design and use AI to drive quality and growth for our clients and our projects in away that enhances our delivery of total design and supports our pursuit of excellence.The AI systems we ‘take, shape, or make’ will accelerate our mission to shape a betterworld, guided by six key principles aligned with our strategy and values. Governance We will Aims This Policy is set by ArupGroup Board andimplemented across all Arupoperations through policy,procedures, managementsystems and learning. Use approved AI systems ina way that is consistent withour aims, other Arup policiesand additional clientrequirements Safety: AI systems operateas we intend them to, withsufficient reliability,robustness and security Compliance: AI systemscomply with relevantregulations and laws Provide our members withthe guidance and trainingthey need to understand AIand action this policy It is reviewed and approvedannually, or more frequentlyif appropriate. Transparency: transparentAI systems foster ourclients’ trust and ourmembers’ confidence inusing AI The Chief InformationOfficer acts for Arup inensuring that plans aredeveloped and implementedto enable us to reach theaims set out in this policywhilst fulfilling applicablelegal, regulatory and otherrequirements. Conduct AI risk assessmentsand implement assurancemechanisms Encourage a culture ofcuriosity and criticalthinking for considered useof AI Accountability: humanshave responsibility for AIsystems and outputs Fairness: AI systems andoutputs do not discriminateunfairly against members,clients, or the public Manage our AI systems overtime to ensure they remainfit for purpose “ It is not the wish toexpand, but the questfor quality which hasbrought us to thisposition, for we haverealised that onlyintimate integration ofthe various parts or thevarious disciplines willproduce the desiredresult.” Bring a diversity of thoughtand expertise to designingAI systems Sustainability:environmental and socialimpacts are carefullyassessed in our use anddesign of AI Handle data for AI systemsethically and securely in linewith our data handlingprotocols Invest in ongoing regulatoryforesight, research, learningand partnerships to maintainbest-practice Jerome FrostCEO, Arup Group LimitedApril 2025 Ove Arup, 1970