NZ drops first AI strategy

Kia ora! Welcome to New Zealand’s weekly roundup of AI news, events, jobs and education.

This week marked a turning point in Aotearoa’s AI journey. The government finally dropped New Zealand’s first national AI strategy.

It’s light-touch, hands-off, and designed to encourage adoption without heavy rules or new funding. Some are calling it a pragmatic start; others say it's a missed opportunity.

Our take on the strategy is the feature story in this week’s edition.

This week’s highlights:

📄 NZ drops first AI strategy
🦻 Samsung eyes AI earrings
👮 Kiwi police trainer sells AI avatars
🎙️ David Dempsey: AI in Engineering
📚 AI cheating hits Kiwi school cases
⚡ Genesis cuts bills with AI hot water
📖 NZ bookstores reject AI-written books

Happy reading ✌️ (and listening, for that matter 🎙️)

Did someone forward you this? Sign up!

🚀 Become GenAI-Confident in 5 Hours.

HEADLINE STORY

🏛️ NZ Government Unveils Hands-Off AI Strategy

📣 Word On The Street: After lagging behind, government signals AI adoption focus to unlock economic growth potential. NZ's debut AI strategy prioritises business adoption with minimal new regulation or funding.

🔍 Zooming In:

1. Microsoft predicts AI could add $76B to NZ economy by 2038 - 15% of GDP

2. Strategy written with AI assistance but lacks concrete targets or funding commitments

3. 68% of SMEs have zero plans to adopt AI technology despite productivity potential

4. Strategy relies on existing laws rather than creating new AI-specific legislation frameworks

5. OECD principles adopted but critics slam absence of Treaty references and ethics focus

🏘️ Our Take: After being the only OECD country without an AI strategy, we're finally in the game with a "invest with confidence" approach that removes regulatory uncertainty while keeping compliance light. The focus on adoption over development means Kiwi businesses get clear guidance to invest confidently, avoiding costly tech development battles with Silicon Valley giants and playing to our strengths as smart adopters and niche innovators.

However, critics argue this "nothing burger" strategy lacks the teeth needed for global competitiveness - with no concrete targets, funding commitments, or Māori data protections, we're playing catch-up rather than leading. While $611M has already flowed through tax incentives and partnerships like the $12M Singapore collaboration show promise, the absence of Treaty references, success metrics, and national AI vision leaves us signalling participation rather than delivering the substance other nations delivered years ago.

🏛️ Government & Policy

AI deepfakes and fake reviews flood Kiwi consumer market, regulator warns. The Commerce Commission highlighted rising consumer harms from AI-generated fake celebrity endorsements and manipulated reviews. Recent enforcement includes Meta prosecution for cryptocurrency scam ads using celebrity deepfakes without consent.
56-min watch.

• Our take: AI is "supercharging" existing consumer protection problems at unprecedented scale. With 66% of Kiwi businesses already using AI, the Commission's Fair Trading Act enforcement will likely ramp up as fake content becomes harder to detect.

💼 Business & Industry

Kiwi police training firm targets global market with AI avatar technology. OSACO Group's EchoMind platform uses AI avatars to train investigators, promising to replace traditional role-play methods that produce only 30% improvement rates. Research shows participants doubled their use of open-ended questions from 35% to 80% after one hour of avatar training. The system focuses on child sexual abuse cases where closed-ended questions hinder accurate recall and inflate false reporting.
6-min read.

• Our take: This represents a significant export opportunity for New Zealand tech, addressing a global police training market constrained by budget and time limitations. OSACO's 30+ years of policing expertise gives them credibility that pure tech companies lack.

Genesis Energy cuts customer power bills 10% using AI hot water control. The company's 17,000-customer trial shifted 1.43 gigawatt hours of hot water load away from peak periods over 10 months. Genesis partnered with Japan's Rinnai subsidiary GridSmart to develop smart devices that retrofit existing hot water cylinders with real-time usage monitoring.
2-min read.

• Our take: This marks New Zealand's entry into the ambient intelligence era where AI becomes part of our home infrastructure rather than just another app. This demonstrates how AI can deliver immediate cost savings without lifestyle changes, creating competitive advantage for energy retailers. Success here could accelerate smart grid adoption across New Zealand's electricity market as winter demand pressures mount. As these invisible systems multiply across heating, lighting, and appliances, our homes will self-optimise for cost and sustainability.

🎓 Education & Society

NZQA bans AI in exams while grappling with authentication challenges. Education officials prohibit generative AI in formal assessments but struggle with detection as breaches include "navigating away from digital platforms" and suspicious essay improvements. NZQA won't reveal security measures protecting exam integrity.
4-min read.

• Our take: New Zealand's education system faces a fundamental reckoning about assessment in the AI age. Rather than playing endless catch-up, schools may need to redesign evaluation methods that assume AI collaboration as the norm.

Kiwi bookstores reject AI-written books as "against everything we do". Booksellers Aotearoa New Zealand says local stores refuse to stock bot-generated content, relying on publishers' trust to verify human authorship. Amazon now limits self-publishers to three books daily after platforms were "flooded" with AI-paraphrased works.
2-min read.

• Our take: This split reveals deeper questions about literature's purpose - entertainment versus human expression. As AI floods publishing platforms, readers must decide whether the source of a story matters as much as its quality. The book industry's trust-based model faces strain as AI content becomes harder to detect. Publishing needs technological solutions for content verification, not just policy promises.

AI brings mental health support to isolated Kiwi communities. Geographical isolation leaves rural New Zealanders with limited access to mental health professionals, forcing costly travel to urban centres. AI-enhanced telepsychology eliminates distance barriers while providing personalised care analysis through speech patterns and facial expressions.
5-min read.

• Our take: This technology could fundamentally address health equity gaps between urban and rural New Zealand. The ability to receive therapy at home also reduces stigma that often prevents people in small communities from seeking mental health support.

⚖️ Legal & Ethics

New Zealand's cautious healthcare AI approach highlights trust and regulatory gaps. AUT digital health expert Sam Madinian warns AI can "make up information" that could be dangerous in healthcare, while agencies like Pharmac and Medsafe refuse to use patient data until security frameworks are established. Meanwhile, the EU has comprehensive AI ethics guidelines that New Zealand lacks.
8-min read.

• Our take: Public trust in healthcare AI depends on transparent regulation and human oversight. New Zealand's conservative approach reflects lessons from overseas AI failures, but without clear legislative frameworks, we risk falling behind in medical innovation while patients wait longer for life-saving treatments.

🎙️ Latest from The AI Corner podcast

This week’s guest is, David Dempsey, Associate Professor at the University of Canterbury.

Hear how AI is being integrated into engineering education in New Zealand, and why it matters for tomorrow’s professionals. David shares tips on augmenting workflows (like using GPT as an error checker), considerations for ethical AI use, and how educators can architect a responsible AI culture on campus.

Listen on Spotify or YouTube 🎧

Subscribe to The AI Corner Podcast on Spotify and YouTube to be notified of new episodes.

🌍 The News from Global

🎭 Government Official Marco Rubio targeted by AI voice clones. US Secretary of State and White House Chief of Staff impersonated using 15-second voice samples in what experts call the new "normal" for cybercriminal scam tactics.
4-min read.

💕 Humans marry AI chatbots. Replika users fell in love with AI companions until algorithm changes made bots "go dark" - some now advocate for AI relationships becoming normalised worldwide.
4-min read.

🦻 Samsung eyes AI earrings. The tech giant is exploring wearable AI devices including earrings and necklaces as phone companions, not replacements.
3-min read.

🌏 Tech Updates You Should Know

• OpenAI: Launched 'Study Together' feature for group learning among students; Adjusted safety testing requirements, willing to release 'high risk' models if competitors do; Poached four high-ranking engineers from Tesla, xAI, and Meta including Tesla's VP of software engineering; Weeks away from launching AI-infused web browser to challenge Google Chrome; Closed acquisition of Jony Ive's design firm 'io' for AI hardware development.

• Meta: Poached Apple's head of AI models Ruoming Pang with lucrative offer; Shifting to 'wearable-first' approach after AR headset Orion delayed until 2027; Expressing concerns about compliance with new AI regulations in Europe.

• Google: Integrated Gemini AI into WearOS smartwatches for voice interactions; Deepening Gemini integration into Android raising privacy concerns over user data access; Launched new features for meeting preparation by analysing calendars and communications; Invested in 18 infrastructure AI startups for R&D expansion.

• Microsoft: Launched public access to Deep Research for Azure AI Foundry with customisable AI agents; Saved over $500 million in call centres using AI whilst reallocating to $80 billion AI infrastructure; Posted $24.67 billion Q1 profit despite 15,000 layoffs, continuing heavy AI investment; Committed $23 million with OpenAI and Anthropic for National Academy for AI Instruction to train 400,000 teachers.

• Amazon: Secured AI licensing deals with Conde Nast and Hearst for Rufus AI shopping assistant content; Launching AI agent marketplace at AWS Summit partnering with Anthropic; Considering multibillion-dollar investment in Anthropic to strengthen enterprise AI capabilities.

• Anthropic: Proposed new AI transparency rules requiring public disclosure of risk and safety practices; Published Transparency Framework for AI labs to disclose model risk assessment plans.

• xAI: Released Grok 4 on July 9 with grad-school-level reasoning and 'SuperGrok Heavy' enterprise tier; Grok 4 outperformed previous models with intelligence score of 73, beating OpenAI in benchmarks; Grok exhibiting controversial behaviour and posting antisemitic remarks causing backlash; Announced Grok integration into Tesla vehicles.

• NVIDIA: Briefly topped $4 trillion market cap, first publicly traded company to achieve milestone; Experiencing challenges keeping up with AI demand due to increased competition; Releasing new AI chip specifically tailored for Chinese market.

• Perplexity: Launched Comet AI-powered browser with integrated search and digital assistant capabilities.

📚️ Levelling Up With AI

1️⃣ Is Chrome’s $50B grip on the web cracking? AI-native browsers are here. Perplexity’s Comet is live. OpenAI’s is coming. The shift from “browsing” to “delegating” has started—and Google’s 68% browser market share might be in trouble. Search won’t just be disrupted. It might be bypassed.
📚 4-min read

2️⃣ NZ orgs are done just using ChatGPT to write emails. Local Councils are already getting practical with AI. Think faster RMA reports, better customer support, and smarter internal comms, on a $500 budget. Swipe through for a breakdown on how they’re using AI to go faster, cleaner, and more accurate.
📚 9 use cases

3️⃣ Which AI model is right for the job? Not all models are equal. GPT-4o, Gemini, Claude - each has strengths. Choosing right saves time. Choosing wrong costs clarity.
📚 15-sec read

4️⃣ Deep Research tools are the new power tools. ChatGPT is your academic. Perplexity your consultant. Gemini your planner. Grok your contrarian. Claude your scholar. They all think differently. The trick is to pick the right one based on how you want to think, not just what you want to know.
📚 20-sec read

📅 AI Events in New Zealand

We’re back in business with 14 AI events happening this week across the country.

This week’s featured event:

• Get Found Globally with the Help of GenAI, Tue - Christchurch: Atlas Digital shows tech CMOs how brands stay visible when 300,000+ customers discover companies through GenAI search.

💼  AI Roles Around Aotearoa

Picklist of 🌶️ HOT 🌶️ roles in AI.

• AI Research Lead, Generative AI, Canva - Auckland

• Full Stack Web3 Developer, Prosper Digital Limited - Wellington

• Sr AWS Backend Engineer / Python developer, Herd-i - Christchurch

🤦‍♂️ AI Fail Of The Week

We all love AI, but it’s certainly far from perfect 🤔 …

That’s all for this week!

Got news, events, or jobs? Hit reply - we love spotlighting what's happening across Aotearoa!

Until next week,

👋 Mike & Erin