Legal expert Damien Steel-Baker provides a bullet point guide to the the EU’s AI Act, coming into force within weeks - and argues that the lack of regulation here could cause uncertainty and inhibit investing in tech that will be governed by so-far-undecided rules. Victoria University’s Dr Andrew Lensen
AI revolution in NZ: Why the speed of human job losses is causing concern
The telco says it’s been investing in AI for three years - it’s just that its efforts have accelerated recently. “We are also actively investing in our people to grow their data and AI capabilities, so that they are prepared to take advantage of the new roles that will come to be through this technology,” a spokeswoman said.
“Overall, we see AI as an enabler of new growth, of efficiency, and of better customer experiences, and we want to help drive its development in New Zealand, so we’re not just a net importer of it.”
AI can create a PowerPoint presentation in seconds, freeing a marketing team for more creative work. It could speed up diagnostic or paper work for a surgeon, giving them more time for life-saving operations.
There’s a slim chance AI could lead to some kind of Terminator-style meltdown, but a much better chance it will help cut hospital waiting lists.
Spark recently showcased local AI innovation in everything from road slip prevention to water quality monitoring. Startups like Prosaic and Zeil are using it to create new businesses much more quickly and easily than before.
The genie can’t be put back in the bottle. The workplace, and society, have been reinvented before.
Lensen doesn’t take issue with any of that. But he does point out an important nuance, and one he thinks our politicians need to take note of.
“People like to use the analogy of the Industrial Revolution, where jobs were lost to automation, but new jobs were created,” Lensen said.
“But what’s different, in this case, is the velocity of it.
“In the Industrial Revolution, machines had to be built to replace people. It took years to make the transition. With artificial intelligence, you know, you can replace humans in a few days if you buy the right product from the right supplier.”
“It’s the short-term harm that I’m most concerned about. We’ve seen companies do massive layoffs this year.
“People might eventually be able to retrain and get other jobs but that’s something that takes time and AI could impact a very large amount of the workforce in a very short period - which is quite distinct from what we’ve seen before.”
So what should be done?
Lensen said the first thing he wanted was for our politicians to simply recognise the problem. In the lead-up to last year’s election, he said politicians were ignoring “the elephant in the room” after reviewing each party’s AI policy - or lack of policy.
Labour promised “a just transition” for AI-affected workers, without describing what that looked like. National said it would ensure AI is used “safely and ethically” – without detailing what this means or how it will be enforced.
Map the problem
The second thing he wants is for the Government to get the lay of the land and fast.
Where are jobs being lost? How many are at risk? And which industries will be hit first?
He wants an inquiry, or at least an investigative exercise.
Action offshore, inaction in NZ
Last month, after a survey found most Kiwis were spooked by AI’s dark side, Technology Minister Judith Collins said: “There will be no extra regulation at this stage”.
That’s a stark difference to the AI Act in the EU, or US President Joe Biden’s series of executive orders on artificial intelligence. Individual US states have also taken steps, such as Utah’s Artificial Intelligence Policy Act - passed this month and coming into force May 1 - which requires transparency when businesses use AI. Should the AI deceive consumers, then businesses could be fined an administrative fine of up to US$2500 ($4100) and/or civil penalties up to US$5000 ($8300).
Collins added: “This Government is committed to getting New Zealand up to speed on AI. We have a cross-party AI caucus, which is due to meet soon. Its first step will be providing feedback on the AI framework we are developing to support responsible and trustworthy AI innovation in government, which the public should expect to hear more on in the coming months.”
She had no update this week.
The EU’s AI Act passed its final legislative hurdle on March 13. Following linguistic checks and other final details, it’s expected to come into force over the next few weeks.
Key banned uses of AI in the EU
Buddle Findlay special counsel Damien Steel-Baker said its key provisions include strict prohibitions on the use of AI for the following, bar some narrow exemptions for law enforcement, if supported by a court, in the case of facial recognition:
- Biometric categorisation using sensitive characteristics such as political, religious or philosophical beliefs or sexual orientation or ethnicity
- Mass scraping of facial images from CCTV footage or the internet
- Social scoring and predictive policing based solely on profiling people (in the vein of Tom Cruise’s Minority Report, based on the Philip K Dick novel)
- Emotion-recognition in the workplace or educational institutions - other than for medical or safety reasons
- Using AI to manipulate human behaviour to circumvent free will
The Brainbox Institute’s Tom Barraclough has pointed out that while New Zealand has no AI-specific laws, artificial intelligence is subject to existing laws including the Privacy Act - which means that, for example, a supermarket has to let its customers know if it’s using facial recognition software (which may or may not have an AI component), take steps to ensure it’s accurate and only be used for the purpose it’s collected (that is safety and theft-prevention, not marketing).
Yet Privacy Commissioner Michael Webster recently called for higher penalties, telling the National Cyber Security Summit in Wellington, earlier this month: “The maximum fine I can issue to an organisation for not adhering to a compliance order is $10,000. Compare that to Australia, where their maximum fine for serious interference with privacy is $50m, and you begin to see the issue.”
Various non-legislative steps are being taken, albeit with different paths being followed by different agencies. Government department Ministry for Business, Innovation and Employment (MBIE) has banned staff from using ChatGPT and its peers altogether, for example, while others are making limited use of AI, through their own guidelines by following recommendations for all-of-Government put together by the Department of Statistics.
But the big picture is that we have no AI laws or formal regulations, and none on the horizon.
Why no regulation could be bad for businesses
At first blush, you might think that’s a friction-free environment, encouraging businesses to experiment with AI.
But Steel-Baker says NZ will adopt AI regulation at some point. The EU legislation will set a benchmark for the future of AI policy here, he believes.
But in the meantime, Steel-Baker says that with no AI laws in NZ, early adopters have to develop systems that may or may not be compliant with future regulations.
“So some will have a bit of uncertainty, which will stifle their innovation.”
Chris Keall is an Auckland-based member of the Herald’s business team. He joined the Herald in 2018 and is the technology editor and a senior business writer.