“Expect AI models to be unaccessible in the EU soon.”
That was the prediction from developer Patrick Senti on X last week, replying to news that the European Union has finally started enforcing its AI Act. The post pulled nearly 100,000 views, and the replies read like a eulogy for European tech ambition: “extremely determined to ossify Europe into a museum.”
The mechanism is familiar and well-documented. When compliance gets expensive enough, providers don’t comply — they cut the jurisdiction off. Meta has done it with its AI tools in the EU. Apple has withheld features. Every new enforcement wave spawns predictions of the next geo-block. Whether this particular prediction lands on schedule or not, the pattern it describes is real: regulation written to protect citizens can end up denying them the tools.
Which raises an interesting question for a small country on the other side of the world: what if New Zealand went the other way — and said so?
New Zealand’s teen bill is not an AI bill
Start with what’s actually on the table here. In late August the Government introduced the Online Safety (Minimum Age and Child Safety Risk Assessment) Bill, requiring platforms like Instagram, TikTok, Snapchat and Facebook to take “reasonable steps” to verify users are over 16 — with penalties reaching 10 percent of global revenue. It’s heavily modelled on Australia’s under-16 social media ban, not on the EU’s AI Act.
And here’s the part that got buried under the “ban the kids” headlines: the bill explicitly excludes AI apps with productivity uses — ChatGPT, Gemini and Copilot are carved out, alongside messaging platforms like WhatsApp. The drafters sorted platforms by risk profile: algorithmic scrolling and endless feeds on one side, tools that help you do things on the other.
So the honest reading of the current NZ framework is this: an under-16 social media restriction aimed at addiction mechanics, with AI productivity tools left alone. Whatever else it is, it isn’t a step toward the EU model. New Zealand’s actual AI policy remains light-touch OECD-style guidance — principles, not prohibitions.
The carve-out is the story
Most coverage treats the AI exemption as a technicality. It’s more than that. It’s a deliberate line in the sand, drawn by a government that has looked at AI and decided these are tools, not threats — the same government, after all, that has been talking up AI for classrooms and public services.
But there’s a difference between a default nobody noticed and a stance somebody claimed. Right now NZ’s openness is an accident of drafting — an exemption in a bill about TikTok. The country is getting the benefit of an open-access AI environment without stating the principle, which means it isn’t defending the principle when it comes under pressure next time.
The honest constraints
Before getting carried away, some realism. New Zealand is 5 million people. No AI provider is going to build a special NZ-optimised compliance path — or spare NZ from a global geo-block decision made for the EU or US market. We mostly inherit whatever larger jurisdictions decide. If a model gets pulled from “the West,” it gets pulled here too.
What NZ can actually do is narrower but real:
- Refuse to add friction. No age verification for AI tools, no licensing regime, no identity-linked access requirements. Every layer of friction a country adds is a layer its citizens pay for.
- Put the onus on the individual, not the platform. Adults are adults. Education over verification — the same logic the Government already uses everywhere else.
- Say it out loud. A stated position — “New Zealand maintains open access to AI tools” — signals to providers and to New Zealanders that this isn’t an oversight waiting to be closed.
Does the teen bill prove NZ just follows orders?
It’s a fair suspicion, but the evidence points the other way — or at least points everywhere except Brussels. The teen bill follows Australia, not the EU. And it may not even survive its own Parliament: NZ First opposes it, with Winston Peters calling Australia’s version a “colossal failure” and warning of a “Soviet-era outcome” for New Zealanders’ privacy. The Greens oppose it from the opposite direction, arguing it doesn’t go far enough in putting duty of care on the platforms. The maths for passage is tight, and the election is in November.
Meanwhile, the substance of the bill — targeting endless scrolling and algorithmic feeds for under-16s — is a defensible position about children, not an opening bid for general internet filtering. The slippery slope is a possibility, not a plan.
The opportunity hiding in the noise
Here’s the pitch: New Zealand could become the country that explicitly chose openness. Not because it’s provocative, but because it’s rational. AI tools are the biggest individual empowerment lever available — they let one person do the work that used to take a team. A country of 5 million can’t out-build China or the United States on models, but it can out-welcome the technology, and small jurisdictions have leveraged exactly that before: light-touch crypto regimes, offshore finance, streamlined company registration. New Zealand has ridden that wave before.
The counter-argument writes itself: safety. But the current NZ government’s own bill answers it — restrict what harms kids, leave the tools alone for everyone else. The “opposite of the EU” isn’t recklessness. It’s the recognition that a compliance regime aimed at frontier labs mostly lands on ordinary users, and ordinary users are the ones who lose when the geo-blocks start.
The EU has made its choice. Australia has made most of its. The United States is mid-argument with itself. New Zealand, for once, hasn’t locked in — and that’s the brief window where a small country gets to pick deliberately instead of by default.
Frequently asked questions
Does New Zealand’s under-16 social media bill restrict AI apps like ChatGPT? No. The bill explicitly excludes AI productivity tools — ChatGPT, Gemini and Copilot are named as excluded, alongside messaging apps. It targets social media platforms with high-risk features like endless scrolling and algorithmic recommendations.
Is New Zealand copying the EU on AI regulation? No. NZ’s AI policy follows the OECD’s light-touch principles model, and its teen online-safety bill is modelled on Australia’s under-16 social media ban, not the EU AI Act.
Could New Zealand formally become an “open AI” jurisdiction? Nothing stops it in law — NZ already has no AI Act, no age verification for AI tools, and no licensing regime. What it lacks is a stated position. Claiming the stance deliberately would protect it from being quietly closed later.
— CJ Murden, editor of Singularity.Kiwi. Former digital technologies teacher, author of AI-focused books. Writing with a New Zealand focus.
