Kia ora! Welcome to New Zealand’s weekly roundup of AI news, events, jobs and education.
Before we get into the news, this is worth a watch. The future of learning is here.
Below is a video of Synthesis 2.0, the world’s first AI math tutor that adapts to each child’s learning style. It’s warm, patient, and interactive, with real-time speech, follow-ups, and visual explanations on demand.
Some will push back on tech replacing teachers, but done right, AI personalises learning so educators can focus on what truly matters.
This week’s highlights:
ChatGPT's drop in market share
Most read: how to stay relevant in AI
Caelan Huntress: hands-on AI fluency
Spark blocks AI child-abuse imagery
Gemini hits 1B visits, shifts web habits
Claude Skills collapse messy workflows
ANZ thwarts $1m fraud with behaviour AI
Tip: ChatGPT’s secret weapon
Japan warns OpenAI over anime use in Sora
Happy reading and listening ✌️
Did someone forward you this? Sign up!
HEADLINE STORY
⚠️ Deepfake porn conviction highlights NZ's outdated laws for AI
Chloe had images stolen from her social media, manipulated and uploaded to porn sites. (Composite image: Vinay Ranchhod) (Source: 1News)
📣 Word On The Street: First-ever NZ prosecution shows how hard it is to nail deepfake criminals when our laws weren't built for AI abuse.
🔎 Zooming In:
Offender used AI to create explicit images from social media photos, shared them widely online.
Victim's IT-savvy dad built fake website trap to identify perpetrator via mobile number.
High school rejection motivated the crime, offender confessed when confronted by victim, and the police secured NZ's first deepfake porn conviction.
🏘️ Our Take: This case exposes a critical flaw in our legal system. Current laws require proof of intent to harm, making most deepfake cases unprosecutable. Unless prosecutors can prove harmful intent, offenders walk free.
With AI tools becoming easier to access, we're seeing a surge in incidents targeting school children. AI-generated abuse material is spreading across mainstream platforms. Schools report rising incidents with some victims attempting suicide.
Act MP Laura McClure's proposed bill could criminalise non-consensual AI manipulation outright, finally protecting us by making it illegal regardless of intent. The 1961 Crimes Act wasn't designed for AI abuse. We need laws that match the technology criminals are using against us. This rare prosecution only succeeded because the victim's dad had IT skills and the offender confessed. Most victims won't be that lucky. The gap between what's technically possible and what's legally actionable is growing every day.
💼 Business & Industry
Spark blocks AI-generated child abuse imagery nationwide. Spark becomes New Zealand's first ISP to block all AI-generated child sexual abuse material, including cartoon-style imagery. The Internet Watch Foundation reports a 400% rise in AI-generated content globally.
2-min read.
Our take: This creates a commercial precedent other ISPs must follow or explain why they won't. Spark's network safety becomes a competitive differentiator in family-focused markets.
ANZ blocked $1M in fraud using AI that watches how you type. AI watches how Kiwis interact with devices to spot scammers in real-time. Reported fraud dropped 9% and customer losses fell 7% in six months.
3-min read.
Our take: This shows how banking security’s moving from reaction to prediction. The AI watches how you type, and catches scammers before they strike. Banking trust translates directly to customer retention and growth.
AI Forum drops governance toolkit that actually works. The refreshed website offers three maturity pathways (AI Ready, AI Smart, AI Trusted) with practical steps matched to align with your organisation's size and risk level.
Link.
Our take: This kills the "framework paralysis" problem. When governance tools are this clear, businesses can move faster with confidence instead of stalling out on theory. Building governance that honours Treaty obligations (which this does) from the start means fairer AI systems for all New Zealanders.
⚖️ Government & Legal
AI is reshaping how directors govern organisations. Boards are adopting AI for meeting prep and risk scanning, but regulators globally are moving fast on AI obligations. Directors face a literacy gap - they need to understand enough to ask the right questions and challenge management.
3-min read.
Our take: AI is creating new categories of director liability. The boards asking hard questions now avoid regulatory pain later. Directors who can't articulate how their organisation governs AI use will struggle to prove they met their duty of care.
Health NZ plans single AI platform for radiology. National RFI targets unified orchestrator amid 7% yearly demand growth and workforce shortages. Platform would route scans to AI models and feed results into existing systems.
2-min read.
Our take: Fragmented pilots don't scale. One national platform could become the blueprint for how public health systems deploy AI without chaos.
🔍 Education & Society
Ministry backtracks on AI marking guidance. Teachers got new AI marking rules this week after last year's total ban. The ministry now says AI can help mark routine work but warns against using it for complex assessments or final NCEA grades.
2-min read.
Our take: This reversal creates a messy grey zone for schools, especially when the Education Minister wants AI marking all NCEA by 2030. The policy zigzag leaves teachers guessing what's actually allowed. The major problem is expecting teachers to become AI experts overnight without training or support. When policy moves faster than capability building, we set people up to fail.
Schools split on AI as cognitive gaps emerge. Wellington East Girls' uses AI-powered podcasts from class slides while Wellington High sees students stressed about falling behind without it. Universities now spot major knowledge gaps between AI-adopter schools and AI-ban schools.
4-min read.
Our take: We're creating a two-tier education system by accident. Without national policy, Kiwi students' critical thinking development depends entirely on which school they attend. This creates workforce readiness gaps that businesses will inherit within 2-3 years.
🤓 Most read article of the week: The Three Paths to Staying Relevant in the Age of AI
AI has flattened the playing field. Some folks are redesigning their roles using AI to multiply their impact. A few are going all in, turning skills into businesses that scale without more hours.
📚️ AI for Business
Helping leaders and teams adapt, learn, and scale with AI.
1️⃣ Claude Skills just launched and I’m completely rebuilding my workflows because of it: What used to take multiple Claude Projects or Custom GPTs and endless copy-pasting now happens in a single, automated flow. This sets the standard for other AI tools to follow.
5-min read.
2️⃣ ChatGPT's market share dropped from 87% to 74% in 12 months: The iron grip is loosening. Only 0.1% of ChatGPT users visit all major GenAI tools, whilst Claude has 27.5% of users visiting 2+ other tools.
1-min read.
3️⃣ Only using Copilot for your AI is like using the internet only to clear emails: Digital natives and startups are experimenting daily with AI, mastering 10-30 tools and wiring intelligence into everything. True fluency looks like connected systems and a mindset built for reinvention.
56-sec watch.
4️⃣ Gemini crossing 1 billion visits in September signals the biggest shift in how we use the internet: Most don't realise what we're witnessing. From searching for links to asking AI for answers to delegating tasks to AI. When Gemini bakes into Search, Android, YouTube, Workspace and Maps, billions use AI without realising it.
1-min read.
5️⃣ We’re witnessing the beginning of the ‘haves’ and ‘have nots’ separation in the AI world: The difference between reading about AI and actually using it will separate the future workforce.
1-min read.
🎙️ The AI Corner Podcast
This week's guest is Caelan Huntress, Head of Learning and Enablement of Agentic Intelligence. Hear:
Why New Zealand was the second most hesitant country in the OECD about AI in 2024, and how government strategy changed that.
Caelan’s hands-on workshop approach where participants use ChatGPT, Claude, and Gemini simultaneously to compare results.
How AI ate up marketing jobs in 2024 as companies realised they could get more from their marketing spend with AI.
💡 Tip of the Week: Custom GPTs Are The Most Valuable ChatGPT Skill To Master
If you use ChatGPT daily, stop prompting from scratch. Custom GPTs turn it into a team of trained specialists that follow your exact process and tone every time.
Why it matters
One-off prompts = one-off results. A Custom GPT gives consistent, repeatable outputs.
You control the voice, structure, and quality.
It scales: share with your team and build systems that run themselves.
How to use it
Go to Explore GPTs → Create → Configure.
Write clear instructions defining the GPT’s role, goal, and success criteria.
Specify tone, format, and examples of what “good” looks like.
Only enable tools (web, code, image, etc.) if they’re essential.
Keep it simple — one GPT, one job.
Think of Custom GPTs as teammates. The best ones have a clear job description, stay in their lane with a narrow skillset, and deliver reliably every time.
Stage 1
Stage 2
🌍 Tech Updates From Global
The selected top headlines from each major AI tech company.
OpenAI
Added parental controls for ChatGPT as a first step to protect younger users from potential AI interaction dangers.
Partnered with Broadcom to develop custom AI chips with a commitment to deliver 10 gigawatts of technology over four years.
Introduced age-verified adult mode allowing erotic content interactions under strict moderation rolling out in December.
Expanded low-cost ChatGPT Go tier to 89 countries beyond the previous 16 Asian markets.
Announced $15 billion investment in AI hub and data centre in India by 2030 to enhance regional AI capabilities.
Introduced real-time AI capabilities into Meet platform and image generation across Search and Lens services.
Launched Veo 3.1 video generation model with enhanced features for video creators including richer audio and better content control.
Anthropic
Introduced Skills for Claude allowing organisations to load specialised capabilities into AI systems across all platforms.
Launched Claude Haiku 4.5 low-cost model with performance approaching Claude 4 and available on free plans.
Entered early talks with Abu Dhabi's MGX for funding shortly after raising $13 billion in Series F round at $183 billion valuation.
Microsoft
Partnered with UK startup Nscale to acquire approximately 200,000 Nvidia GB300 GPUs for Azure cloud infrastructure.
Upgraded Windows 11's Copilot into an AI agent capable of automating various tasks to enhance user productivity.
Announced $3 billion investment in cloud and AI infrastructure in India to establish regional presence.
Amazon
Launched AgentCore AI agent platform aimed at providing businesses with AI-powered solutions for enterprise needs.
Meta
Announced construction of gigawatt-sized data centre in El Paso, Texas with $1.5 billion investment for AI infrastructure expansion.
Started using Arm's energy-efficient Neoverse GPUs for certain AI workloads to enhance performance and efficiency.
Reached 1 billion monthly users of Meta AI with standalone app demonstrating significant user adoption.
Apple
Launched new MacBook, iPad and Vision Pro devices powered by M5 chip featuring significant GPU and AI performance upgrades.
Testing AI-driven Siri version to integrate more seamlessly with operating systems and diminish reliance on traditional app icons.
Miscellaneous
Japan's government formally urged OpenAI to cease using copyrighted manga and anime imagery in video app Sora.
Walmart formed partnership allowing customers to link accounts and purchase items directly through ChatGPT feature.
📅 AI Events in New Zealand
20 events this week!
This week’s featured event:
The Power of Artificial Intelligence in Reducing Access Barriers, Wed - Christchurch: Access Advisors CEO, Dr Chandra Harrison will show how AI tools are breaking down digital barriers for disabled and neurodivergent Kiwis, helping execs design with inclusion in mind.
📅 Promote your event with us. Reply to let us know.
💼 AI Roles Around Aotearoa
Picklist of 🌶️ HOT 🌶️ new roles in AI this week.
AI & Automation Specialist, Ryman Healthcare: Christchurch
💼 Promote your job with us. Reply to let us know.
🤦 ️ AI Fail Of The Week
We all love AI, but it’s certainly far from perfect 🤔 …
👋 Mike & Erin