AI in September

Check out the latest AI learning opportunities, a CROSS safety report on unverified AI-generated structural calculations, major model developments and perspectives on how AI can strengthen, rather than shortcut, engineering judgement.

Upcoming AI webinars and workshops

Engineering New Zealand webinars and workshops

Explore practical ways AI can help engineers deliver better projects. Join Tim Greene and Aaron Johnson from Caliber Design for our next monthly webinar on Wednesday 21 October, from 12.00pm to 1.00pm. They will share practical examples of how engineers are using AI to organise project files, automate repetitive CAD tasks and connect engineering analysis with report preparation. The webinar will also provide a framework for identifying where AI could support your own work, whether you are new to AI tools or already exploring workflow automation.

Watch AI as a productivity tool in an electricity utility. In this Electrical Engineering Group webinar, Dan Martin shares practical examples of using AI for analysis, coding, reporting and asset management. He also discusses the governance, risks and professional responsibilities engineers need to consider when using AI in practice.

Exploring the Power of Generative AI. Our next full-day, in-person workshop is being held in Queenstown on Thursday 12 November. Register to build your understanding of generative AI and explore how it can support your engineering work.

AI Forum New Zealand update

AEC Knowledge Hub. The AI Forum's Architecture, Engineering and Construction Working Group continues to add practical material for people working across engineering and the built environment.

Tool Deep Dive: Copilot for M365 in AEC Practicec. Join Maria Mingallon online on Wednesday 21 October, from 12pm – 1pm, for a practical look at where Microsoft 365 Copilot can support reports, specifications, correspondence, meetings and project data, and where human judgement remains essential.

AI in the Built Environment. This free in-person event in Auckland on Wednesday 4 November will share case studies from water, transport, energy and the built environment. The focus is on how organisations are using AI to improve decisions, unlock value from data and address infrastructure challenges.

Potholes in the AI Expressway: the complete series. The AEC Working Group's series brings together industry leaders to discuss the opportunities, risks and practical realities of AI adoption across the built environment. It covers practical use cases, data sovereignty, governance, professional accountability, commercial and insurance issues, and lessons from earlier industry transformation.

Aotearoa AI Summit 2026 highlights. Nearly 600 people attended this year's two-day summit in Wellington. The AI Forum's September update recaps the event and highlights follow-up work on accessibility, the AI Blueprint for Aotearoa and the 2026 Aotearoa AI Awards.

Engineering and Aotearoa related AI developments 

Designing effective collaboration with AI. How do we design working experiences that help people collaborate effectively with AI? A new University of Auckland white paper, Working with AI: A Design Framework for Human-AI Collaboration, brings together engineering, social and psychological perspectives to explore this question.

The report offers six design principles and a five-step process for applying them. A human–robot assembly case study illustrates how choices about communication, automation and human control can affect efficiency, trust and the effort required to work with a system. It provides a practical starting point for engineers introducing AI into workplaces, with attention to both performance and people’s experience.

Spark scales data-centre connectivity across Auckland. Spark has announced a dedicated network connecting Auckland data centres at speeds from 10 to 800 gigabits per second. The new network will provide faster, more resilient connections between Auckland data centres to support growing cloud and AI workloads.

AI system helps laboratory devices ‘talk’ to each other. A new platform allows laboratory instruments from different manufacturers to communicate and be controlled by an AI agent. The approach could help researchers automate experiments without replacing their existing equipment.

AI Day 2026: Engineering with AI. The University of Waikato's third AI Day will be held in Hamilton on Thursday 22 October, alongside the Waikato Engineering Design Show. The programme will examine practical and agentic AI across design, analysis, modelling, coding, project delivery and asset management, as well as professional accountability, governance, data security and human oversight.

Inappropriate use of artificial intelligence in structural design. A CROSS safety report describes an architect relying on AI-generated calculations to justify heavy plant on an existing roof. Although the report looked detailed and credible, a structural engineer found that its technical content was fundamentally incorrect and the roof was only capable of supporting light plant. The case is a clear reminder that plausible presentation is not evidence of technical accuracy. AI-generated engineering information must remain within professional competence and be checked by suitably qualified engineers before it is relied on.

AI trends to watch

Architectural visualisation with GPT-6 Astra. An OpenAI developer used GPT-6 Astra through Codex to turn a simple house brief into an editable 3D scene in Blender, then developed it into an Unreal Engine walkthrough. It is an early example of AI moving beyond text and images to operate professional software and help build spatial, editable outputs.

AI infrastructure investment could reach US$31.6 trillion by 2050. PWC projects that annual global data centre investment could rise from about US$800 billion in 2026 to US$1.8 trillion in 2050. Power availability is expected to be the main constraint, making AI infrastructure an increasingly important issue for the energy, water, planning and built environment sectors.

Interesting reading

Introducing System One Models and Jev. TypeSafe AI is proposing a different model architecture for decisions embedded in software. Instead of generating free-form text, Jev returns predefined, typed outputs with probabilities and confidence scores. The company's claims still need independent testing, but the approach is an interesting example of AI being designed for fast, structured automation rather than chat.

When AI builds itself. Anthropic reports that, as of May 2026, Claude authored more than 80 percent of the code merged into its codebase. The article explores how AI is accelerating its own development, while also noting that substantial gaps remain in choosing goals, exercising broader judgement and reviewing work as quickly as AI can generate it.

From our AI Advisory Committee

Robotics on Richard's radar

Those who have been reading the AI update for a while, are aware of my fascination with humanoid robots.

• Atlas vs Optimus is a great discussion of the two paths developers have taken: a traditional robotics approach built on decades of engineering knowledge, and a newer AI-first approach. As a former automation engineer it is great to see that knowledge of things such as inverse kinematics appears to be winning.
• Why Boston Dynamics Gave Up on Humanoids focuses on the hardware and engineering behind humanoid robots. Despite the title, it's less about abandoning humanoids and more about the design decisions, components and trade-offs involved in building them.
• What is the state of art in drones? Some of the Cargo Drones are remarkable.

Richard Templer is the CE of Engineering New Zealand Te Ao Rangahau and Chair of the Engineering New Zealand AI Advisory Committee.

Member spotlight: Maria Mingallon

“If you are using AI mainly to shortcut the work, you may also be shortcutting the learning that makes your work valuable.”

Maria Mingallon is AI Leader for Asia Pacific, New Zealand and Australia at Mott MacDonald, a member of Engineering NZ AI Advisory Committee, Chair of AI Forum New Zealand, and Chair of its Architecture, Engineering and Construction Working Group. She advises organisations on critical technology and AI challenges that can reshape how they operate, make decisions and develop their people.

In Better engineers, not just faster engineering, Maria argues that the real opportunity for AI in engineering is not simply to automate more work or produce answers more quickly. It is to help engineers become sharper critical thinkers by making accumulated experience, organisational knowledge and the reasoning behind past decisions easier to access, question and improve.

Hear why she thinks the real test of AI is not whether it helps engineers reach an answer faster, but whether it helps them ask better questions and form judgements they can defend.

Read Maria's perspective

Let us know what you think, what you’d like to see more of, or share a success story of your own – get in touch at ai@engineeringnz.org