Kia ora! Welcome to New Zealand’s weekly roundup of AI news, events, jobs and education.
A short read this week during a quiet period for the AI ecosystem locally and internationally. The highlight from Big Tech for the holiday season: higher usage limits!
Merry Christmas and a happy New Year all.
This week’s highlights:
NZ leads world in AI skepticism
One NZ rolls out AI customer service
Kiwi actors fight AI voice cloning globally
ChatGPT launches year-in-review feature
Anthropic doubles Pro limits for holidays
Microsoft launches Copilot Eggnog Mode
OpenAI plans advertising rollout for 2026
Google makes Gemini 3 Flash the default
Apple delays personalised Siri to 2026
Happy reading and listening ✌️
Did someone forward you this? Sign up!
💼 Business & Industry
AI buoy gives Golden Bay whale rescuers a head start. Solar-powered tech now monitors Farewell Spit 24/7, detecting pilot whale distress calls and alerting responders in real time. The system could expand to shipping ports, feeding whale location data to operators to prevent collisions.
2-min read.
Our take: The commercial potential is intriguing, if proven at scale, this tech could become standard equipment for any industry operating near marine mammals. Port operators will be watching closely.
Your telco calls now answered by AI. One NZ rolled out AI across customer service and networks, with CTO sharing hybrid human-AI model at global Mobile World Congress 2025.
4-min watch.
Our take: We're already talking to AI when we ring our telco. The question isn't if this happens, but whether it actually improves service.
We're the world's most AI-skeptical nation. NZ leads global markets in AI wariness: 62% fear losing human connection, 66% demand human service options, 45% distrust government AI use.
2-min read.
Our take: Trust gap threatens competitive advantage. Companies need transparent AI strategies and upskilling plans before technology deployment.
Kiwi actors fighting AI voice cloning globally. Jennifer Ward-Lealand's new role tackles unauthorised AI use. French actor found his synthesised voice at train stations without consent or payment.
3-min watch.
Our take: If your voice or likeness can be cloned and monetised without permission, this fight matters to everyone, not just performers.
🌍 Tech Updates From Global
The selected top headlines from each major AI tech company.
OpenAI
"Your Year with ChatGPT" Wrapped-style feature launched in NZ/US/UK/CA/AU with personalised awards and AI-generated poems; ChatGPT Atlas security update with adversarially-trained model against prompt injection; Codex 2× holiday usage boost through Jan 1; Formatting Blocks UI for in-chat document editing rolling out to Plus/Pro; Advertising plans reported with sidebar sponsored content mockups for H1 2026; AI Foundations Certification Program launched with Indeed/Upwork job platform integration; Chain-of-thought monitorability research published showing longer reasoning improves misbehaviour detection
Anthropic
Holiday 2× usage limits for Pro/Max through Dec 31 across web/Claude Code/Chrome; BLOOM agentic safety evaluation framework open-sourced under MIT with 0.86 Spearman correlation to human judgment; Copyright lawsuit filed by Pulitzer winners opting out of $1.5B September settlement
xAI
US Department of War partnership for GenAI.mil serving 3M personnel at IL5 security clearance by early 2026; Grok Collections RAG API launched with free first week indexing and $2.50/1K searches; Grok image editing on X launched without creator consent sparking artist exodus; Grok now powers full X algorithm analysing 100M+ posts daily; Colossus 2 "MACROHARD" expansion with 230K GPUs operational, targeting 50M H100-equivalent GPUs and 2GW power, raising $20B
Microsoft
Copilot Eggnog Mode holiday feature released; Defender Experts Suite announced with managed XDR/threat hunting/incident response GA Jan 1 2026; 2025 Copilot Usage Report published analysing 37.5M conversations; Windows AI rewrite denial after employee LinkedIn post caused outrage; GitHub Copilot WRAP methodology published (Write/Refine/Atomic/Pair)
Google/DeepMind
Gemini 3 Flash confirmed as default globally in Gemini app and AI Mode; Free US users get "Thinking (3 Pro)" reasoning mode with daily limits; AI Pro 50% off promotion through Dec 31 at $9.99/month; Research breakthroughs review published by Dean/Hassabis/Manyika covering AAIF and MCP; Musk publicly backs Hassabis over LeCun on AGI debate
Apple
Personalised Siri confirmed delayed to iOS 26.4 in 2026; AI Chief Giannandrea retiring 2026, Amar Subramanya hired from Google Gemini team; China Apple Intelligence must refuse 95% of 2,000 government test prompts partnering with Alibaba Qwen3; iOS 26 Wallet AI order tracking launched extracting from emails without third-party integration; $12.7B capex vs $380B combined competitors noted in year-end analysis
Meta AI
AI chatbot interactions now personalise ads globally (excluding EU/UK/South Korea) since Dec 16; Italy Competition Authority suspends WhatsApp rival AI chatbot ban as abuse of dominance, Meta appealing; AI moderation flagged 635K+ accounts in 2025 with legitimate business false positives reported; "Avocado" frontier model in development at TBD Lab under Alexandr Wang for text/coding
Perplexity
2025 Review reflection tool launched with preset prompts for year-end AI summaries; Patents beta continued promotion as first AI patent research agent; Amazon lawsuit context covered re: Comet browser agentic purchases, Morgan Stanley predicts 50% US shoppers using AI agents by 2030
🤦 ️ AI Funny Of The Week
We all love AI, but it’s certainly far from perfect …
👋 Mike & Erin