Online Media Regulation - Continued

The Pause That Refreshes

On Friday 15 May 2026, after a debate in Parliament, Minister Erica Stanford announced that the Bill introduced that would restrict access by under 16’s to certain social media platforms would be “paused”.

There was some discussion about whether or not this could in fact be done unilaterally. The Bill is on the agenda and to stop Parliament voting on the bill, the Government would have to rely on Opposition MPs agreeing to an 11th hour change to Parliament’s agenda.

But the reason for the “pause” is interesting and concerning. Stanford’s office said:

“The Government is undertaking a wider programme of work in this area and Wedd’s bill is on hold for the moment. We are continuing to work through the process and will have more to say soon.”

Ms. Stanford has made it clear that the Government would have more to say “next month” (June). She said

“The Government is steadily progressing with work on social media and online harm policy, and due to this, Catherine Wedd’s Member’s Bill is being deprioritised in place of wider work”

This, of course, follows the Education and Workplace Select Committee report which made recommendations that were substantially wider than an under 16 social media restriction.

Key recommendations included:

1. Independent national online safety regulator (top priority) The committee’s most structurally significant recommendation was the establishment of an independent national online safety regulator. The body would hold a broad mandate to respond to emerging harms, develop secondary legislation, and serve as a single public point of contact. The committee estimated it could be operational within an existing agency in the near term, with full independence achievable within two to three years.

2. Social media ban for under-16s A majority of committee members favoured a social media ban for under-16s, similar to the one introduced in Australia. The committee acknowledged age verification challenges and that some young users would find workarounds, but concluded “the level of online harm currently being sustained by young people warrants introducing an age-delay restriction as part of a broader reform agenda.” The Greens and ACT dissented.

3. Strengthened liability for platform design harms The committee recommended strengthening liability for online harm arising from platform design features, such as the use of algorithms and infinite scroll features.

4. Ban on “nudify” deepfake applications An immediate ban on “nudify” applications that automate the creation of deepfake sexual imagery was recommended.

5. Algorithmic restrictions for young users The committee called for restrictions on algorithmic recommendation systems for young users.

6. Advertising regulation The committee recommended regulation of online advertising for alcohol, tobacco, and gambling products targeting under-18s.

7. Research and education The committee recommended New Zealand-based research and its promotion, as well as educating and empowering parents, caregivers, and young people. Teachers also called for professional development to support the digital tools being used in schools.

Overall framing: The committee framed New Zealand’s approach as that of a “fast-follower” rather than a “first mover,” noting that both lagging behind other countries and trying to be unique could result in less effective outcomes. The report concluded that current law is not adequate and called for “a proactive, system-wide approach to social media platforms.”

During the debate on 13 May 2026 Stanford said a ban was no “silver bullet”. But she said it would help change how people used social media.

“It takes time, but it does change social norms. But it is not the silver bullet on its own, because, I’ll tell you what, the social media giants would love a ban because, as you know, children can try and get around it.

That’s why it is so very important to have a strong regulator that changes their behaviours. That is going to be the key,”

So Ms Stanford favours a strong regulator to change the behaviour of social media giants – and also internet users.

That must send something of a shiver up the spine. Clearly Ms. Stanford is into control.

Some weeks ago I published a proposal for media regulation. It can be found here.

That proposal was made available to the Coalition partners. Given that National seems to be driving the media regulation issue I would have expected to hear something in response – even it it were to acknowledge the work that I had done. Nothing.

I wonder if Ms. Stanford is starting from scratch or whether she is going to provide us with something similar to the Safer Online Services and Web Platforms proposals of the Department of Internal Affairs under the Ardern\Hipkins Government. One would hope that it is not the latter.

But I have recast my proposals. What I first published was a formal proposal. I have decided to simplify the proposal – make it more readable and perhaps more accessible.

I recommend it to those in Wellington who seek a solution. Seek no further. The solution is here. Give me a call.

In the meantime here is the simplified version.

How New Zealand’s Proposed Media Regulator Would Work

A plain-language guide to the Media and Communications Authority proposal

New Zealand’s media landscape has a problem that most people who use the internet can feel, even if they cannot quite name it. A teenager bullied through Instagram has one place to turn. A reader misled by a news website has another. A viewer offended by a streaming service has almost nowhere to go at all. And the platforms that algorithmically amplify harmful content — that design their systems in ways that, often by intent, keep people enraged and engaged — face no coherent regulatory accountability in New Zealand whatsoever.

The result is a regulatory architecture that is fragmented, incomplete, and increasingly inadequate for the world it is meant to govern. What I propose in response is something considerably more ambitious than a patch on existing law: a unified Media and Communications Authority (MCA) that would consolidate the functions of several existing institutions under one roof, bringing news standards, professional content regulation, and platform harm oversight into a coherent relationship for the first time.

Understanding how this authority would actually work requires examining both its foundations — how it defines harm, and why — and its structure: three distinct divisions, each serving a different part of the media landscape, all bound by common principles and supported by shared institutions.

The Problem with What We Have Now

New Zealand currently manages media and communications through four main bodies, none of which was designed with the contemporary digital environment in mind.

The Harmful Digital Communications Act 2015 (HDCA) and its administering agency Netsafe handle individual complaints about harmful digital communications. The Broadcasting Standards Authority (BSA) covers linear broadcast television and radio, but expressly excludes on-demand content — the very medium through which most New Zealanders now consume audiovisual content. The New Zealand Media Council (NZMC) offers voluntary self-regulation for news publishers, but has no statutory powers to enforce its decisions. And the Office of Film and Literature Classification operates a prior restraint system for classifiable publications, a regime rooted in a very different era of media distribution.

The most significant gap is in platform regulation. Content that spreads harm through social media or algorithmic recommendation systems falls between these institutional jurisdictions.

A newspaper that publishes a defamatory story can be brought before the Media Council. A social media platform that algorithmically amplifies it to hundreds of thousands of people faces no equivalent accountability at all.

The HDCA comes closest to bridging this gap, but my analysis finds it structurally unfit for the task. Its definition of harm — “serious emotional distress” — is narrow and subjective, excluding reputational damage, economic harm, threats to physical safety, and broader dignitary injuries.

More critically, it was designed for interpersonal communications between individuals. It has no tools to address the systemic harms that arise from platform design, recommender systems, and the architecture of algorithmic amplification.

It places the full burden of enforcement on individual victims, who must prove personal harm in order to trigger any regulatory response. And it was written before the full implications of AI-generated content, volumetric online harassment, and platform-mediated radicalisation had become apparent.

The HDCA remains necessary but cannot be sufficient.

Starting with Harm

Before building a regulatory structure, the proposal establishes what that structure is meant to prevent. This is not a trivial exercise. “Safety” — the organising concept most often invoked in online regulation discussions — is, an elusive basis for regulatory intervention.

It is capacious enough to justify almost anything, and therefore disciplined enough to justify almost nothing.

The proposal instead centres its analysis on harm: demonstrable adverse effects on physical safety, mental health, wellbeing, dignity, or fundamental rights. The definition is deliberately constrained. Harm must be more than trivial, and must be supported by empirical, clinical, or reasonably inferable evidence. Crucially, the definition expressly excludes mere offence, discomfort, disagreement, or exposure to controversial or unpopular ideas. Being upset by a viewpoint is not harm. Being harassed into clinical anxiety may be.

The definition also extends the concept of harm beyond individual items of content to encompass the effects of platform design itself — algorithmic amplification, recommender system failures, dark patterns in user interface design. This is the conceptual move that transforms the framework from a reactive complaints mechanism into something capable of addressing systemic platform accountability.

This definition threads a constitutional needle that matters enormously in a liberal democracy. Section 14 of the New Zealand Bill of Rights Act guarantees freedom of expression. Section 5 permits that freedom to be limited only by reasonable limits demonstrably justified in a free and democratic society.

A harm definition that sweeps in mere offence would fail that proportionality test. A definition too narrow to capture systemic platform harms would fail the purpose of regulation altogether. The proposal claims to sit in the productive middle ground.

The Three-Part Framework

On top of this definitional foundation, a hybrid regulatory model is proposed built from three structural elements.

First, codes of practice. Industry participants — social media companies, streaming services, news publishers — develop codes that set standards for managing harmful content and platform design.

These codes are reviewed and registered by the regulator, but their content is determined by industry, not government. Once registered, compliant platforms receive a “safe harbour” — protection from certain civil liability. This creates a strong commercial incentive for participation without imposing direct state control over content.

The insistence on state separation here is deliberate and emphatic. No government department may play any role in determining the content of codes. The proposal’s language on this point is notably firm: any perception of state involvement in content standards risks transforming the regulator into what may be called “a quasi-censorship arm of the State.”

The model draws on co-regulatory experience from Australia and the United Kingdom, while explicitly rejecting their more interventionist features.

Second, a reactive-harm complaints mechanism. For cases where harm has occurred to an individual or group, an accessible complaints pathway exists. Netsafe — the existing online safety organisation — remains as a triage and mediation agency.

Where Netsafe cannot resolve a complaint, it is referred to an independent Communications Tribunal for determination. The Tribunal’s decisions can be appealed to the High Court on questions of law.

Third, a backstop regulatory power. Where industry genuinely fails to develop adequate codes, or where compliance is insufficient, the regulator can impose mandatory standards.

This power is explicitly a last resort, available only where self-regulation has demonstrably failed, and any standards issued under it must satisfy the same freedom of expression proportionality analysis as industry-developed codes.

One Authority, Three Divisions

Rather than creating three separate regulatory bodies to administer these functions — each with its own board, budget, infrastructure, and governance burden — the proposal collapses them into a single institution. The Media and Communications Authority operates through three distinct but complementary operational divisions.

Division 1 — News Media Standards is the successor to the New Zealand Media Council and absorbs the BSA’s jurisdiction over news and current affairs content.

Its scope is defined by function rather than technology: any publisher — print, broadcast, digital, podcast, or online-only — that exercises a genuine fourth-estate function falls within its ambit.

Membership is entirely voluntary, but the incentives attached to membership are substantial: qualified privilege in defamation proceedings, Privacy Act exemptions, journalist source protection, eligibility for New Zealand on Air funding, and an MCA quality mark.

The proposal’s reasoning is that any serious news organisation will find these benefits sufficiently valuable that voluntariness does not undermine meaningful coverage.

Division 2 — Content Standards fills a gap that currently has no institutional home at all.

Internet-only broadcasters, on-demand audiovisual services, and streaming platforms with a New Zealand business presence are currently subject to almost no content regulation. Division 2 creates a voluntary co-regulatory framework for these providers, with industry-developed codes reviewed and registered by the division.

A minimum revenue threshold of NZD 500,000 determines which providers are formally invited to participate; smaller providers can opt in voluntarily.

Global streaming platforms without a New Zealand business presence — Netflix, Amazon Prime, and their equivalents — are expressly excluded, acknowledging the practical enforcement limitations of domestic regulation over multinational digital services.

Division 3 — Online Harm is the most structurally novel element and the most consequential. It addresses the platform harm regulation gap directly, combining the extended HDCA framework with the code-compliance model described above.

Participation in code development is compulsory for “responsible platforms” — a defined category covering social media, content-sharing services, search engines, and messaging services with a material New Zealand user base.

Code compliance itself remains voluntary, but non-compliance forfeits safe harbour protection.

Netsafe continues as the triage and mediation agency for individual complaints, with unresolved matters escalating to the Communications Tribunal.

The compulsory participation element of Division 3 creates a potential tension with the voluntary character of Divisions 1 and 2. The proposal addresses this explicitly, requiring the governing legislation to state clearly that participation in one division does not affect obligations under another, and that Divisions 1 and 2 remain voluntary regardless of how Division 3 is structured.

Keeping the Regulator Independent

The institutional design is constructed around several layers of independence, each addressing a distinct failure mode.

The MCA’s governing board is appointed by an independent appointments panel rather than by government. No current employee, officer, director, or shareholder of any regulated entity may serve. No current or recent government official may serve.

Board members must have collective expertise spanning law and constitutional rights, journalism, digital technology, and consumer protection. They serve fixed terms and can be removed only for specified cause. This architecture is designed to resist both regulatory capture by industry and political capture by government.

The Communications Tribunal, critically, sits outside the MCA entirely. It is established as a fully independent adjudicative body, receiving matters referred from all three divisions and from the amended HDCA.

Its independence from the MCA is structural: the regulator, which may be engaged in code development and industry standard-setting, should have no influence over the body that adjudicates compliance with those standards. Tribunal decisions are published in full. Appeal to the High Court is available on questions of law.

Netsafe continues to operate independently of the MCA, performing its triage and mediation functions under the amended HDCA. The legislation would confirm that Netsafe cannot be a government department and cannot be directed by the MCA.

The BSA is retained in a transitional capacity, narrowed to entertainment content on linear broadcast platforms while its news and current affairs functions transfer to Division 1 and its online and on-demand functions transfer to Division 2. The long-term legislative trajectory is for the BSA’s residual jurisdiction to be absorbed into Division 2 by amendment.

The Office of Film and Literature Classification is not affected by the framework at all. Its prior restraint model, appropriate for the most serious categories of objectionable content, continues as a separate instrument.

The Legislative Architecture

The proposed framework requires four pieces of legislation. The principal statute — the Media and Communications Authority Act — establishes the institution, its three divisions, the governing board, the Communications Tribunal, the code development processes, the backstop powers, and the key statutory definitions of “news media” and “responsible platform.” The Harmful Digital Communications Amendment Act updates the HDCA’s harm definition, extends platform-level liability, and connects Netsafe’s triage function to the Tribunal. A Broadcasting Amendment Act makes the transitional adjustments to the BSA’s jurisdiction. And a Consequential Amendments Act updates the Privacy Act, Defamation Act, Official Information Act, and other affected statutes to reflect the new definitional and institutional landscape.

The principal Act must also contain explicit prohibitions preventing any minister or government department from directing the MCA about code content, complaint outcomes, or regulatory discretions. This is described in the proposal as “the most important feature of the entire framework” — a judgement grounded in the comparative experience of jurisdictions where government influence over media regulators, however indirect, has been used to chill expression and compromise institutional integrity.

The Risks, Honestly Stated

The proposal has the unusual virtue of identifying its own principal risks explicitly, rather than leaving them to critics.

The first is contamination between divisions — the risk that the compulsory character of Division 3 will be perceived as undermining the voluntary character of Divisions 1 and 2, making news media regulation appear government-backed by association. The mitigation is explicit statutory separation and clear public communication, but the risk cannot be entirely designed away: institutional proximity creates perceptual proximity, whatever the legislation says.

The second is board capture. A single board governing all three divisions is a single point of institutional failure. The appointment criteria and independent appointments process provide structural protection, but history in comparable jurisdictions offers grounds for measured scepticism about how reliably such protections hold over time.

The third is governance overload. Three operationally distinct regulatory regimes under one board represents a substantial governance burden. The mitigation — Divisional Directors with substantial operational autonomy, with the board focusing on policy and accountability rather than operational detail — is sensible, but requires ongoing institutional discipline to maintain.

The proposal acknowledges that a three-body framework remains a viable alternative if the government or media industry concludes that the institutional separation of news media standards from platform harm regulation is essential to sustaining public confidence in the voluntariness of Division 1.

What the Framework Is Not

There are limits to what is being proposed.

This is not a censorship framework. The foundational commitment throughout is to freedom of expression as the default. The regulator cannot tell publishers what to publish or platforms what content to permit. Its authority runs to the processes by which harm is managed and to the systems platforms use to amplify or mitigate it — not to the editorial decisions of publishers or the content choices of users.

This is not a solution to every problem the internet creates. Unlike broader “online safety” proposals, the framework is deliberately anchored to harm as its organising concept.

It will not prevent harm from occurring. It will not make the internet comfortable or frictionless. What it offers is a proportionate, principled mechanism for mitigation — one calibrated to New Zealand’s constitutional requirements and institutional scale, rather than imported wholesale from larger jurisdictions with different regulatory contexts and market conditions.

And it is not a framework built from scratch. It deliberately preserves and builds on existing institutional infrastructure: the BSA’s four decades of broadcast standards expertise, Netsafe’s experience with online harm complaints, and the voluntary commitment of news organisations to the Media Council’s self-regulatory tradition. The architecture is designed for the New Zealand that exists, not a hypothetical jurisdiction of unlimited regulatory appetite and enforcement capacity.

The Central Bet

At its core, the proposal places a bet on co-regulation: the proposition that industry, given clear frameworks, genuine incentives, and a credible backstop threat, will develop more effective and more legitimate content standards than any government agency could impose directly.

It is a bet informed by comparative experience — the Australian and British models have demonstrated both what co-regulation can achieve and where it tends to fail — and bounded by hard constitutional constraints that New Zealand’s Bill of Rights framework makes non-negotiable.

Whether that bet pays off depends less on the elegance of the institutional design than on the quality of the codes that industry develops within it, the independence that the Communications Tribunal maintains in practice, and the political will to resist the pressure — which will come from multiple directions, as it always does — to use a media regulator for purposes other than the ones it was designed to serve.

The proposal provides a framework. What it cannot provide is a guarantee that the framework will be used as intended. That, in the end, is the irreducible condition of any regulatory design that takes freedom of expression seriously.