This is the third article in my series which considers the proposals of Dr Gavin Ellis set out in his position paper entitled “If not journalists, then who”. This article considers Dr Ellis’s proposals and lines them up alongside those of the Law Commission. I consider the solution proposed in that light and offer some proposals of my own for the future of media regulation
Returning to the Proposals
Let us revisit the proposals of Dr. Ellis and line them up alongside those of the Law Commission from 2013. My comments are in italics.
Firstly it is recognized that any changes should not be approached on an ad hoc basis but have a vision “predicated on a healthy, sustainable, pluralistic system that provides democratically significant journalism and promotes social cohesion across the breadth of society.”
It is noted that the objective of social cohesion features prominently and I restate my position that what should be preferred is the concept of social stability.
The proposals then suggest that there should be a series of inter-related legislative/regulatory changes that represent those areas where Government can make material differences to the sector’s future. The first – overcoming the market-destabilising and distorting effect of transnational digital platform operations in New Zealand.
Clearly it is contemplated that Government action is needed to “level the playing field” based on market distortion from the operation of transnational platforms. This form of State intervention itself is a form of market distortion because it is an attempt to create an artificial outcome. One possible solution that seems to appeal is that proposed by the Fair Digital News Bargaining Bill.
Funding is critical and that form of Government based artificial support underpins four priorities that Dr Ellis identifies:
1. Address systemic effects of medium-specific law and regulation on structures and functions, including the oversight of content standards.
2. Change commercial and taxation law to facilitate innovative and sustainable media ownership structures.
3. Enact laws that protect against detrimental effects of artificial intelligence while promoting its benefits.
4. Redress the imbalance, for practical purposes, in legal liability between news media and social media in relation to reputation, privacy, and harmful content.
The oversight of content standards is concerning although if what is suggested is the Code based model that is currently employed by the Broadcasting Standards Authority there can be little cause for concern. The BSA model is triggered by a complaint so is retrospective in nature whereas a content control model that involves pre-publication control is prospective. This model is employed by the Classification Office which may censor material because it is objectionable or may impose viewing guidelines. If this model were to remain it is essentially the status quo. Given that Dr. Ellis’ later recommendation is to fold the Classification Office into the overall regulator, care would have to be employed to ensure that there was no extension of existing powers.
Changes in commercial and taxation law are interesting.
Legislating for the detrimental effects of AI is a recipe for failure. To start with AI is a term that involves a number of different computational activities. At the time of writing Large Language Models like ChatGPT are front of mind but are but one type of AI. Predictive coding, document prioritization or Technology Assisted Review (TAR) involve elements of AI in the litigation discovery process. This is a beneficial use of AI but has been employed in the litigation process for a number of years. One wonders if legislation would be required for this aspect of AI. Similarly machine learning and development is present in a number of areas of activity and one wonders about the wisdom of addressing these. Finally the whole field, in common with other aspects of the Digital Paradigm, is subject to continuing disruptive change. By the time Parliament gets around to legislating the field will have moved on and AI may, like blockchain, be yesterday’s new new thing.
Finally as far as imbalance between news media and social media are concerned in relation to reputation, privacy and harmful content these elements are already covered by present law. The law of defamation has met the challenge of online reputational harm and indeed some of the early cases about the internet involved actions in defamation. Privacy is clearly covered by the Privacy Act and harmful content is covered by the Harmful Digital Communications Act. It is difficult to understand the concern about any perceived imbalance.
What is further proposed is a co-ordinated and coherent approach which could be met by a single agency “having a mandate to recommend to Government a comprehensive suite of legislative changes that will reset the media landscape for the 21st century.”
In some respects this echoes the Law Commission proposals.
The paper then goes on to state preferred options which I transcribe verbatim with some minor adjustments.
Preferred Options
Government should initiate a comprehensive, coordinated review of legislation governing regulation of both state-owned and private sector media. Although the Broadcasting Act may form the kernel from which wider legislative change can flow, it should not be seen in isolation from other necessary legislative reforms.
For that reason, only the immediate need for platform neutrality in the Television New Zealand and Radio New Zealand Acts, and the consequences of changes to distribution formats should be considered in advance of addressing the following:
• Opening discussions with media operators with a view to creating a new independent regulator, based on (suitably updated) Law Commission recommendations, which adopts existing standards pending agreement on unified terms and standards.
• Scoping the merger of the activities of the Classifications Office into the single media regulatory body.
Following those discussions we consider the best option for coordinating regulatory and other media-related functions is through [a single entity]. We believe the future status of the Advertising Standards Authority should be left to a later stage, when other reforms have been completed.
There are other areas that require expedient review.
The case for a single independent body to oversee media regulation was proposed by the Law Commission over ten years ago. There really is nothing new here as far as the general approach is concerned. It is interesting that the Classification Office may merge with this body. This introduces a content control element to the proposal. And it also opens the door for mission creep. Even a separate body such as the Classification Office is prone to this. During the pandemic the then Chief Censor David Shanks advocated an extension of the powers of the Office to address what he perceived to be online harms. Thankfully that discussion went no further but, as I shall later discuss, it has not completely died.
There may be a case for proposing such a merger but given that the Classification Office operates under the Films Videos and Publications Classification Act 1993, operating under carefully defined statutory powers in what is a censorship role and where an established jurisprudence has developed I have some concerns with such a merger. Better, in my view, that a body who role is that of censorship should stand apart from the activities of mainstream media.
Artificial intelligence
Media organisations seek urgent protection against the use of their content to ‘fuel’ generative AI.
Government should consider investigating amendment to the Copyright Act as an interim form of protection of intellectual property rights relating to news media content (including archives) that may be subject to Large Language Model scraping.
The legal implications of AI also need urgent study to:
• Ensure news media intellectual property is protected.
• Ensure mātauranga Māori, te ao Māori, and the cultural treasures (including language) of Pacifika peoples are protected.
• Avoid the unintended consequence of hindering legitimate news media innovation using AI.
The issue of AI is a highly complex one. Defining what is meant by AI is a problem in itself. AI is notoriously difficult to define, and there is probably no single satisfactory definition that most researchers would agree to. This proposal focusses on generative AI and Large Language Models (LLMs) such as ChatGPT or Microsoft CoPilot.
LLMs are trained on large-scale text corpora, which are collections of texts from various sources and domains. The size and quality of the text corpora are crucial for the performance and generalization of LLMs. The larger and more diverse the text corpora, the more likely the LLMs can capture the linguistic and factual knowledge of natural language and handle different types of inputs and outputs.
The text corpora used by LLMs are often scraped from the internet, which is a rich and dynamic source of information. For example, GPT-3, one of the largest and most powerful LLMs, is trained on a text corpus called Common Crawl, which contains about 45 terabytes of text data from over 60 billion web pages. Other sources of text data include Wikipedia, news articles, books, social media posts, blogs, forums, and reviews. Some LLMs also use domain-specific text corpora, such as scientific papers, medical records, legal documents, or customer service transcripts, to fine-tune their models for specific tasks or applications.
From this brief background the intellectual property (IP) implications for MSM outlets using the Internet becomes very clear. The problem will be which end of the process should be approached. It seems to be suggested that the acquisition of content should be the point of approach. How one crafts a rule that provides for the protection of MSM IP against scraping to populate the Common Crawl will in itself be a challenge. It may require a paradigm shift in IP law theory. The reason that I say that is this. IP law is based – at least in the copyright area – not on the acquisition of an idea but upon the expression of an idea. If I express an idea in a particular way, I own the copyright or the intellectual property not to the idea but to the expression of the idea. If someone without my authorisation takes my expression of an idea, uses it and monetizes it that person has infringed my copyright. Up until the point of use of the expression of the idea there is no infringement. A person is fully entitled to read the book or article where I expressed the idea. There is no prohibition on the acquisition of the expression of the idea (unless it is the result of unlawful copying or other form of piracy but lets not get too complicated)
So armed with that theory scraping data in and of itself may not be a form of infringement. Infringement depends on the use of the data. It is here that we have a problem because IP law recognizes that there should be fair use of another persons expression of an idea. Fair use does not extend to the reproduction of the entire work but it allows for a segment to be used – say in a review or in an academic article where the work is quoted.
The other problem is the way that the scraped data is used. If it is directly copied and published in a way that goes beyond fair use there could well be infringement. But if it is paraphrased or precised there may well not be an infringement.
And as if the problem was not complex enough we must remember that the output of LLMs depends upon a “prompt” or query from a user. And variation in prompts may well result in a variation of responses from the LLM.
So the issue of IP protection from AI scraping and for mātauranga Māori, te ao Māori, and the cultural treasures (assuming there is protectable IP in these areas) is a highly complex and nuanced exercise that will probably result in a paradigmatically different approach to IP law.
Reputation, privacy, and harmful content
Government should undertake comprehensive review of laws relating to reputation, privacy, and harmful content to determine:
• Individual and collective fitness for purpose, including:
• Impact of common law
• Social media liability
• Appropriate forms of redress
• Their ability to accommodate rapidly changing technologies.
It seems to me that this element has been included in the options in the interests of completeness. The law relating to reputational harms, especially in the area of online defamation is well developed, clear and mature. There seems to be little need to revisit the Defamation Act. Similarly the liability of social media platforms in the area of reputational harm is well settled and needs little if an adjustment.
The Privacy Act was amended in 2020 and appears to be fit for purpose and the issue of online harms is covered – at least for individuals – by the Harmful Digital Communications Act 2015.
If there is any area that needs adjustment it lies in the complex area of how the New Zealand courts assert jurisdiction over off-shore based platforms. This is especially problematic in the area of the Harmful Digital Communications area and in my view is a matter that requires review.
Funding
Options
Some direct funding options and structures have been discussed already in addressing the levelling of the playing field. However, changing circumstances will also necessitate a reappraisal of the ways in which the sector receives indirect forms of financial support.
That reappraisal will necessarily include the ways in which NZ On Air operates, including changes to its funding models.
We also see real merit in providing structures that reflect the public good contributions that some forms of news media struggle to maintain.
We urge Government to consider:
• Recognising certain forms of journalism as a charitable purpose, and enabling trusts to nominate journalism as their primary purpose.
• Investigating the creation of tax-free business entities modelled on the U.S. Low Profit Limited Liability Company but with direct eligibility for news media
Extreme care needs to be employed in considering the provision of funding for MSM by the State. Any funding provision must be at an extended arms length from the State and could well be provided by the establishment or recognition of vehicles which allow a better funding model for MSM.
In my view recognizing a charitable purpose for forms of journalism, enabling trusts to nominate journalism as a primary purpose and the use of tax relief or tax free entities provides degrees of separation that were not present in the Public Interest Journalism fund.
Structure
Options
The following preferred options are predicated on agreement being reached with private sector media – principally but not limited to those currently members of the Media Council.
We believe strongly that the fragmented and often anachronistic state of New Zealand news media will only be fully remedied by a unified approach under a single umbrella.
For that reason we urge Government to consider reconstituting the Broadcasting Commission as a commission responsible for the administration of the following autonomous, independent functions:
• Content standards and complaint adjudication
• Content classification
• Funding allocation
Research and advocacy would be a responsibility of the commission.
As it appears to function effectively and remains fit-for-purpose, we believe it would be preferable to maintain the Advertising Standards Authority as a stand-alone self-regulator until all other stages of reform are completed, at which point negotiations should begin with affected parties with a view to integration.
This harks back to the Law Commission proposals of 2013 and perpetuates the single regulator model that is preferred.
Care must be employed in considering content regulation. What is proposed is of limited scope, employing a standards-based approach employed by both the Broadcasting Standards Authority and the New Zealand Media Council. Necessarily there would have to be a separate set of standards for broadcast and for “print” MSM but this could be easily accomplished by a single regulator.
What must not happen is for the regulator to be involved in decisions about content control and classification (as suggested). In my view, and for reasons that I have stated above, that should remain with a separate and independent Classifications Authority
In Conclusion
The media landscape and the Internet ecosystem have undergone major changes since the Law Commission Report of 2013. Yet it is interesting that the basic model proposed by Dr Ellis does not differ that markedly from what was suggested eleven years ago.
The preferred options and recommendations are very general in their language and do not go into great detail as one might expect from what is essentially a high level document.
In my view the some additions may well provide a model that will enhance public trust and confidence in MSM, that will enable MSM to be an objective and trusted disseminator of matters of fact and opinion for the citizenry, that will allow MSM to “speak truth to power” and hold the powerful or the corrupt to account and thus provide a trusted informational foundation for social stability that allows for a diversity of opinions and views.
The Single Regulator
- Should encompass the roles of the Broadcasting Standards Authority and the New Zealand Media Council
- Membership of the Regulator will automatically provide members with the legal privileges and responsibilities of journalists
- Membership of the Regulator should be voluntary as proposed by the Law Commission
- Membership of the Regulator should be open to Online Platforms and providers
- Members of the Regulator will be bound by Codes of Practice set by the Regulator and compliance with those Codes will be contractual
- The Regulator will have the ability to entertain and adjudicate upon complaints brought by members of the public against members.
- The Regulator will have no role in content control which will be in the hands of the Classification Office which will continue to operate under the provisions of the Films Videos and Publications Classification Act
One thing that will NOT enhance public confidence in these proposals is if they are allowed to become some form of ersatz Safer Online Services and Media Platforms regime. Those proposals were wide ranging and intrusive and impinged considerably upon freedom of expression and the robust nature of discussion that characterizes the online space.
Although the Safer Online Services proposal has been shelved there is still a significant amount of support within certain sectors of the community and some civil society groups. This article from the CEO of InternetNZ is an example. The hunger for Internet Regulation is still present. The proposals by Dr Ellis in my view and with some modifications go as far as necessary and no further in the regulation of entities that use Internet-based communications.
social cohesion - the very word gives me the shudders. It may be a tired on account overuse label, but to me much of what Ellis proposes is in Orwell’s 1984 territory. I admire Ellis but my caveat is that he has spent a long time ‘in the tent’. The effect, as with the judiciary, is to become myopic in relation to the wider world.