Artificial intelligence has been relegated to the too-hard basket this election. Allyn Robins from The Brainbox Institute argues that’s a big mistake. 

It may well be that 2023 is looked on as the year AI finally went mainstream. Since the release of ChatGPT late last year, AI has become the focus of unprecedented academic, commercial, and media scrutiny. Some predict it will usher in an unparalleled age of prosperity, others fret it may cause the extinction of the human race. Still others, such as Brainbox, argue that it will simply be a very powerful and delicate tool.

But what almost everyone can agree on is that AI matters, and that this is a very important time for AI. What is done now by researchers, businesses, and governments will determine the trajectory of arguably the most significant technology of the 21st century.

Yet despite the topic’s near universally-acknowledged importance and with election day looming, AI is conspicuously absent from the policy platforms and websites of New Zealand’s major political parties. Despite a brief spat over National’s use of AI-generated images in political advertising, few political figures have been willing to take a public stance on what, if anything, a new government should do about the technology. If pressed, most will (quite fairly) point out that the topic is a complex one, and that few politicians truly understand it. Many will add that they’re reluctant to impose regulations for fear of “stifling innovation,” or that legislation is futile because “the technology is moving too quickly to regulate.” Across the board, AI regulation seems to have been relegated to the too-hard basket.

But effective AI regulation is one of the most impactful and forward-looking things an incoming government could offer Aotearoa. AI tools are already entrenching bias, undermining privacy, and enabling non-consensual sexual imagery, including of children and young people. These harms affect New Zealanders already, and if left unchecked will only grow in the future.

If politicians are worried about stifling innovation, they should heed recent Ipsos polling showing New Zealanders consistently report lower understanding of and less trust in AI than the rest of the world; a recipe for stymied innovation if there ever was one, and an issue legislation could make great strides in addressing. Confidence and stability are essential for sustainable innovation, and private industry is not equipped to provide them alone, especially when it comes to a technology that’s generated so much hyperbole, dread, and legal action.

International legislators and industry leaders alike agree that sensible regulation will enable innovation rather than stifling it, by increasing trust in the technology and its implementations, aligning our domestic expectations with a growing consensus in international markets, and by giving commercial entities confidence that their products and/or experiments are conducted with the blessing of the law rather than merely its ignorance.

Equally importantly, regulation can ensure that innovation is focused on areas that actually benefit New Zealand and its citizens. Cryptocurrency is an area that saw a huge amount of “innovation” while it was unregulated, but largely in creating new and exciting ways to conduct fraud. That’s a kind of innovation we can do without.

And the task of regulating AI doesn’t need to be as daunting as it sounds. Yes, the technology is extremely complex and fast-moving, and trying to regulate individual techniques or implementations is almost certainly a losing proposition. But focusing on what people do with AI rather than the nuts-and-bolts of the technology is perfectly viable.

Recognising this, there is no shortage of internationally-recognised examples and frameworks that lawmakers can look to, from the EU’s landmark AI Act to the OECD’s AI principles. Rather than getting bogged down in the minute technical details of AI systems, these efforts are based on principles of accountability, transparency, non-discrimination, and human oversight, ensuring that they will remain relevant and flexible as the technology develops. 

There is a clear way forward on AI regulation that builds on well-accepted fundamental principles about what should guide AI development and deployment. And while the public service is working to grapple with this foundation, we haven’t seen the same engagement from elected leaders. That lack of engagement is dangerous, and it will have consequences.

Regulation does not need to be immediate, or perfect, but it is necessary to ensure that New Zealanders can experiment, adopt, and interact with AI safely and confidently. With many forms of AI just now reaching mass adoption, the choices made by Parliament in the next few years will determine the relationship that Aotearoa and its people will have with the technology. It’s worth putting in the work to make sure it’s a good one.