AI in August

AI New Zealand

• Engineering New Zealand AI webinars – watch our August webinar Designing 3D pipe networks with AI and look out for our second Panel Q&A on AI and engineering education.
• Victoria University guidance – Te Herenga Waka has released new advice on student use of AI.
• University of Auckland TeachWell hub – refreshed hub resources and a practical action plan for generative AI in teaching; new coursework guidance helps lecturers integrate AI responsibly.
• Royal Society Te Apārangi – best-practice guidelines on generative AI in research.

AI in action

• AI Forum productivity report – the third national report highlights productivity gains across sectors, but warns of skills shortages, adoption challenges, and governance. Local case studies show how organisations are improving efficiency and sustainability.
• AEC working group – sector-specific collaboration exploring practical AI adoption for architecture, engineering and construction. Early outputs include shared case studies and pilot frameworks.

Learning and events

• Aotearoa AI Summit – the flagship event for AI in Aotearoa brings together leaders from industry, government and research to showcase real-world adoption, policy shifts and cutting-edge applications. Expect keynote insights, sector-focused sessions, and plenty of networking with decision makers and practitioners driving AI’s future in New Zealand.
• M2 AI Summit – focused on business adoption, this summit highlights how organisations across Aotearoa are moving from pilots to scale. Hear case studies from local enterprises, explore practical frameworks for ROI, and connect with peers tackling the same challenges of workforce, governance and implementation.
• AI workshop for engineers – our full-day generative AI course has achieved a world-class Net Promoter Score of 71. New course and webinar dates for September to November will be announced soon.
• This Is AI for Business – six-part programme for Kiwi business leaders, powered by ANZ and NewZealand.AI. From AI basics to building your digital workforce, all sessions are complimentary.

AI New Zealand trailblazers

• AwareGroup + Boston Dynamics – a rare collaboration between a Kiwi AI development firm and Boston Dynamics bringing advanced robotics, including the Spot quadruped robot, into New Zealand operations. The partnership is showcasing how AI-driven robotics can add value in real industrial settings and even new media contexts.
• MAUI63.org – this conservation project uses UAVs equipped with AI to track rare marine mammals around Aotearoa’s coasts. The approach provides real-time insights that help protect vulnerable species while reducing the cost and risk of traditional monitoring methods.
• Fruitometry – a New Zealand company applying AI-driven yield modelling to transform orchard management in New Zealand. By giving growers more accurate data on crop size and quality, the technology supports smarter decision-making and more sustainable production.

AI Global

• AI technical standard – the Australian Government has published a new AI technical standard to guide the safe, transparent and responsible design of AI systems. It sets out principles for developers and organisations to align with international best practice
• Colab AutoReview – Colab has launched AutoReview, an AI design review tool. It could speed up technical checks in engineering workflows while reducing iteration time.
• Matilda model– Australian researchers and industry partners have launched Matilda, a locally built large language model aimed at providing a sovereign alternative to global AI platforms. The project highlights growing regional interest in developing independent AI capabilities
• Hardware efficiency – researchers have developed an AI system that can rapidly identify and design new catalysts, potentially accelerating breakthroughs in clean energy and sustainable manufacturing. The approach combines machine learning with quantum simulations, cutting discovery times from years to days
• Atlas robot progress – Toyota and Boston Dynamics revealed major upgrades to the Atlas humanoid. The new version demonstrates greater agility and control, marking real progress in robotics.
• Mass intelligence is here – Ethan Mollick argues GPT-5 signals the start of “Mass Intelligence,” where advanced AI becomes ubiquitous. Like search engines once were, it could soon be a universal layer across engineering projects.
• Novo Nordisk lessons – MIT Sloan Management Review shares governance and culture insights from a large Copilot rollout at Novo Nordisk. Key takeaways include deliberate change management, clear guardrails and measuring outcomes, not just feature adoption.

AI tools and software advancements

• GPT-5  – launched Thursday 7 August, promising stronger reasoning, coding and multimodal input. The release has drawn criticism.
• Microsoft GPT5 rollout – Microsoft has integrated GPT-5 into 365 Copilot, GitHub Copilot, Azure AI Foundry and Copilot Studio. A new “smart mode” enables more dynamic reasoning and response optimisation.
• Apple Intelligence meets GPT5  – Apple confirmed GPT-5 will power its intelligence framework in iOS 26, iPadOS 26 and macOS Tahoe 26. This brings more advanced AI features directly into Apple’s productivity and design tools.
• AutoCAD 2025 – The latest release introduces Autodesk Assistant and AI-driven drafting features. These aim to speed up repetitive design tasks and improve accuracy for engineers and architects.
• Codex with VS Code – OpenAI’s coding agent, Codex, now runs on GPT-5 within Visual Studio Code. This expands real-time code generation and debugging support for developers.
• NotebookLM – Google has expanded video overviews to 80+ languages and improved audio summaries. A redesigned Studio panel makes multitasking easier, building on July’s Canvas and PDF upload features.

Security Spotlight

• Agentic browser risk – Comet’s indirect prompt-injection flaw. Brave warns that Perplexity’s Comet browser treats webpage content as trusted, allowing hidden instructions to hijack user sessions. The remedy remains incomplete and traditional web security measures (like same-origin policies) no longer apply.
• Scale AI leaks confidential docs via public Google Docs – Business Insider uncovered that Scale AI left thousands of files publicly accessible via Google Docs links, including sensitive materials for Big Tech clients (Google, Meta, xAI) and personal contractor data. The company has since disabled public sharing and launched an investigation.
• OpenAI Safety Tests – Anthropic and OpenAI collaborated in early summer 2025 to run misalignment-focused safety evaluations on one another’s models and have now published the results.
• Safety and security – Agentic browsing raises fresh risks from indirect prompt injection. Brave’s audit of Perplexity’s Comet browser is a useful case study, and OpenAI and Anthropic have been publishing more about evaluations and auditing agents.

Give it a go

• Google Cloud’s AI emissions calculator – explore the energy and water footprint of your AI prompts with Google’s new tool. It’s a practical way to see the hidden costs of model use.
• Prompting GPT-5 effectively – this podcast episode shares 11 actionable techniques for steering GPT-5 more reliably. This is useful for anyone refining their day-to-day workflow with AI tools.

We encourage you to share your AI insights, case studies or resources that could benefit the wider engineering community – or let us know what resources you would like to see provided.

Get in touch