Published on the 11/06/2024 | Written by Heather Wright

AIRA calls for action to avoid falling to global competition..

The creation of programs to encourage private sector adoption of AI, government leadership in adoption of the technology, and an ‘effective’ national data infrastructure are required to enable New Zealand to reap full returns from AI, including export of the technology.

A whitepaper by the Artificial Intelligence Researchers Association (AIRA) warns that a lack of investment in AI research and adoption would see the country lagging behind and unable to compete effectively with dominate platform firms domiciled in large markets.

“Supporting local AI research and development rather than buying solutions from overseas will improve the social benefits of broader national AI adoption.”

“This situation risks Aotearoa New Zealand losing the ability to tailor AI to our local needs, priorities and ethical standards, and not being independent in terms of technology or data sovereignty,” the whitepaper, which was authored by academics from the universities of Waikato, Canterbury, Auckland and Wellington, says.

“The main message of this whitepaper is that we must invest in leveraging our strong AI research base to increase the competitiveness and productivity of Aotearoa New Zealand in a manner that suits our needs and priorities.

“Building and strengthening mechanisms the efficiently disseminate AI expertise from research to industry has the potential to significantly increase productivity and prosperity for all of Aotearoa New Zealand.”

The paper says AI research will be critical in ensuring the creation of AI imbued with characteristics and values important to Kiwis, while recognising the multicultural nature of the country and te ao Māori values.

“The decisions that AI systems will make are a reflection of their creator’s values and beliefs. This is why it is so important when it comes to AI research, to design systems that prioritise local values and ways of being, and. Not solely commercial goals.”

But it’s also important to ensure profits from AI innovations remain in New Zealand, and reduce money flowing to AI solutions from overseas, improving the social benefits from broader national AI adoption, and to bolster New Zealand’s woeful productivity.

“Innovation in AI is key to lifting productivity by automating tasks with low and high economic value.”

The whitepaper is the second edition following on from a 2021 offering.

Three years down the track, AIRA’s vision and recommendations remain the same, including increased funding for public AI research by developing a network of research centres, hubs and programs in basic and applied AI research, and talent development.

The development of a taskforce to establish standards and regulations for the ethical use of AI, including culturally appropriate AI and use of AI to promote socially inclusive growth are also among the key recommendations.

While the years since the 2021 release have seen rapid advancements in AI, including the advent and advancement of large language models, they’ve also seen declining fortune for New Zealand in the AI stakes. The Global AI Index which ranks nations based on investment, innovation and implementation of AI, has New Zealand in 36th position, a significant drop from the 22nd place it held in 2019, and well behind Australia, at 15. The United States, China and Singapore head the list.

AIRA’s call for investment comes with a warning, noting that that investment needs to be strategic, focusing AI ecosystems around New Zealand’s vital sectors to avoid spreading small amounts of resources across every sector and in doing so failing to see greatest return on investment.

It has identified four focus areas: Sustainable AI, responsible foundations for AI, increasing investment from government and industry, and AI education in terms of competence and expertise.

Within that AIRA says primary industry, climate change and environment, health, high value manufacturing and social and ethical considerations are the five sectors of social application and national priorities where AI can play a role in addressing national challenges.

“AI has the potential to be used in our primary industry [including dairy, aquaculture, open ocean, viticulture, animal products, forestry, horticulture and plant and animal disease diagnosis], climate change mitigation, health sector and high value industry, as well as social/ethical considerations to improve our economy and establish our AI economy.

“The government has started paying attention and making investments in these areas, but there is still a long way to go – and great potential for impact.”

‘Substantial’ investments are required to enhance fundamental research and apply the techniques, the report says.

The country already has key strengths in many of the areas touted in the report, including AI/ML abilities which could make ‘significant’ contributions to the primary sector.

The report says New Zealand also has several key areas of strength which could provide ‘immediate and great help’ for the high-value manufacturing industry, including cybersecurity, sustainable/renewable energy, material and chemical engineering, robotics and drones, and IoT and robotics.

The report is also pushing for a unified and sustainable regulatory environment that clearly explains how AI is implemented and how data can be collected, stored, processed, shared, used and potentially deleted.

While other countries are adopting both soft and hard regulatory stances, New Zealand has lagged. An Algorithm Charter was developed several years ago, requiring fair, ethical and transparent use of algorithms by government agencies. But the report notes ‘transparency appears to be quite weak compared with other international practices’.

Earlier this year, Science, Technology and Innovation Minister Judith Collins said the government was gearing up to develop AI regulations. A cross-party group is looking at a draft framework for AI use in government.

Creating programs to encourage private sector adoption of AI is a key part of the paper, with AIRA recommending the creation of an AI ecosystem with industry labs, and the creation of AI digital innovation hubs to drive the technology into SMEs. Funding for AI startups is also recommended.

The need for a good data infrastructure in order for AI to be successful is also called out. The AIRA recommendations include the creation of an effective national data infrastructure with open data partnerships and datasets while enabling and supporting Māori data sovereignty obligations. The creation of test environments and regulatory sandboxes is also called for.

On the talent side, it’s calling for new initiatives to help students and the overall workforce develop the required skills, and the provision of better work conditions to attract the best professors and scientists to New Zealand.

They also include a doubling of the number of researchers, lecturers, senior lecturers, associate professors and professors in AI by 2027 and again by 2033, with similar increases in students studying AI, increased government funding for AI research, as seen in Europe, and investment in the Strategic Science Investment Fund to enhance CRI-university collaborations.

“If Aotearoa New Zealand does not invest in research, AI will only be efficient software running in the cloud of large overseas companies. This outcome creates risk and could negatively impact our country’s independence in terms of technology and data sovereignty.”