News Item

An intriguing new report, co-authored by Tom Barraclough, Hamish Fraser and Curtis Barnes and published by Brainbox Institute, warns that enacting legislation in code is a risk to the constitutional arrangements of New Zealand. Despite that, they argue that two pieces of legislation currently under review should be drafted in computer code: the Resource Management Act 1991 and internet filtering legislation (currently at Select Committee).

Increasing use of computer systems to give effect to law and policy

Domestically and globally, governments are grappling with how to use regulation to deal with digital policy issues. They are also increasingly using computer systems to give effect to regulatory systems. This report was prompted by a global movement initially led by New Zealand’s innovative “better rules approach”. Law as code has subsequently been explored internationally by “rules as code” advocates including in Canada, Australia, Singapore and the United Kingdom. The topic is a revival of research with a history that began before the development of the modern internet.

Passing legislation in code creates constitutional risk

The authors find serious reasons for concern about the idea of Parliament passing “legislation as code”. Enacting legislation in code undermines the separation of powers and the role of the Judiciary. It also confers too much power on Executive government. They say “legislation” should not be passed in code.

The researchers argue that while law as code is the future, it must remain subordinate to written law.

Computational models of statutes can be useful

By contrast, they say computational models of the law can be a useful tool to improve government service delivery, access to the law and access to justice.

Greater parliamentary guidance required for automated decision-making

The authors identify laws which already empower government agencies to use “automated electronic systems” to make decisions and perform other legal tasks. They say that Parliament should give greater guidance as to how the “reasonable reliability” of such systems is to be assessed, including whether the interpretation of the law being implemented in these systems is legally reliable.

Policy development should account for digital systems implementation

While code should never be legislation, they say the “better rules approach” to policy development pioneered in New Zealand carries significant promise in any situation where law might be implemented in computer systems. The approach uses a multidisciplinary team to take a “service design” approach to how policy will be implemented, vastly improving the practical workability of legislation in digital systems.

Resource Management Act should be re-drafted using “law as code” approaches

In the future, we will want to process resource consent applications as much as possible using computer systems. On that basis, we should re-draft the Resource Management Act using a better rules approach.

Internet filters legislation is hazardous

The authors also apply their findings to a current legislative proposal to implement a nation-wide compulsory internet filter to prevent access to objectionable material. They say the legislation delegates too much power to Government agencies to dictate how that “electronic system” would work, noting that review and appeal mechanisms to challenge the filter have not even been written yet for public scrutiny.

Incubator or sandbox should be founded to fuel innovation and build trust

The public and others have good reason to be skeptical of “law as code” initiatives. To build domestic expertise, as well as public trust and confidence, the authors recommend an incubator or “sandbox” be founded to foster further development. This would allow government, academia, the private sector, and civil society groups to create transparent computational models of the law together.

By working in the open, digital systems can be scrutinized and assessed for their lawfulness and the way they will be deployed before they are implemented. Building this expertise could have significant developments for upskilling lawyers and computer scientists, generating exportable commercial opportunities, and improving government policy development and service delivery.

A full copy of the report is available here

The Law Foundation provided funding of $101,521 for this research through its Information Law and Policy Project.