Artificial Intelligence – Friend or Foe?
The first event in our Speaker’s Science Forum series for 2024 was held on 20 Poutū-te-rangi March at Parliament. Professor Alistair Knott discussed what it would take to keep AI safe; and Professor Richard Green presented an array of AI applications in conservation and agriculture.
Keeping AI safe, making AI useful: Two snapshots from the AI policy frontlines
Professor Alistair Knott, Te Herenga Waka—Victoria University of Wellington
Professor Knott talked about how the Web, and the world beyond it, are about to be flooded by a wave of AI-generated content. We need reliable ways of identifying such content, to supplement the many existing social institutions that enable trust between people and organisations. This is important because:
- AI-generated content undermines accountability of human organisations such as companies or universities
- AI content generators threaten to destabilise information ecosystems, because individuals can generate much more content that ever before.
Professor Knott also discussed provisions in the EU's AI Act, and in President Biden's recent Executive Order on AI, which should result in better detectors for AI-generated content becoming available. Policymaking for large AI providers is best tackled internationally and it is imperative that New Zealand participates in these discussions. Similarly, we need good public evaluation frameworks for the AI systems that are deployed locally.
For New Zealand to benefit from the productivity and profits that AI offers, we need NZ-based AI companies, and these companies need skilled engineers. We are at risk of missing opportunities because of capability issues that arise from the loss of AI staff from our universities.
AI applications to see the unseen in conservation and agriculture
Professor Richard Green, Te Whare Wānanga o Waitaha University of Canterbury
Professor Richard Green, Te Whare Wānanga o Waitaha University of Canterbury, described the recent explosion of AI accuracy, processing speed and cost savings and how they have suddenly enabled applications undreamt of even five years ago – especially for challenging outdoor applications in conservation and agriculture.
He presented diverse projects covering real-world applications including drones for pruning forests and safely managing vegetation around powerlines, which protects human operators and staff in this high-risk industry. For one example, his team used a deep learning approach to recognise trees, deploying cameras on existing rubbish truck networks to automatically scan areas around ports of entry for biosecurity threats and incursions.
Professor Green also described the autonomous underwater vehicles his team have developed, which are able to inspect mussel lines for yield or health issues, detect invasive biofouling species on wharf pylons, scan for holes in salmon nets for aquaculture, and map the seabed to locate scallops for precision harvesting with minimal environmental disruption.
Overall, Professor Green’s work clearly demonstrated that AI has reached the stage where we can automate agriculture and aquaculture and with this technology, we can improve primary production, accelerate sustainability, food security and biosecurity.