Cabinet Minister Judith Collins wants the government to expand the use of artificial intelligence (AI), starting with the health and education sectors where it could be used to assess mammogram results and provide AI tutors for children.
“It doesn’t do the work for them. It says some things like ‘go back, rethink that one, look at that number,’ those sorts of things. What an exciting way to do your homework if you’re a child.”
Deploying AI in education and health would be seen as high risk uses under new legislation passed by the European Union regulating AI.
Using AI in those settings in EU countries must include high levels of transparency, accuracy and human oversight.
But New Zealand has no specific AI regulation and Collins is keen to get productivity gains from extending its use across government, including using it to process Official Information Act requests.
An OIA request by RNZ for a government Cabinet paper on AI was turned down (by a human) on the grounds that the policy is under live consideration.
Collins told RNZ she already uses ChatGPT to write drafts of her speeches.
She seems exactly the type. There’s something about AI enthusiasts I really don’t like, and I find it difficult to put into words, but it’s a certain combination of undeserved optimism and for the technology, an unshakeable belief that it will solve all our problems, and perhaps the most infuriating thing, a blindness to just how fucking weird AI responses typically are.
And given that Judith is probably a lizard person in a skin suit, it doesn’t surprise me she’s blind to AI weirdness.
In the survey, released last month, 69 percent of New Zealanders said they had a good understanding of AI and 64 percent thought it would profoundly change their lives in the next three to five years.
I know that The Dunning-Kruger effect has been disproven; but fuck a “good understanding” means what exactly? The people I work with (are they representative of the population? Maybe…) have trouble using excel well, most people have no idea how their computer works or why it works. Just because you can use something does not mean you understand it.
Work paid for me to go to a “getting started with AI for businesses” seminar run by [redacted reputable organisation name] and holy crap the FOMO.
- The whole premise of the thing basically boiled down to “LLMs are a massive game changing technology that is going to make huge amounts of human tasks obsolete and if you don’t get in on it now your competitors will and you’ll be bankrupt in a decade” which… idk. Useful technology for sure, but this isn’t the AI singularity. The vibe I got was all these people are old enough to see the fortunes won and lost when the internet exploded, and are terrified that this is going to be that all over again and that they’ll end up left behind.
- People massively personify LLMs without thinking through the actual detail in how they work. Someone asked a question about how you can rely on information the LLM gives you, and the suggestion was to just ask it how confident it is which isn’t really how LLMs work - they are fancy auto complete, it has no theory of mind or actual reasoning - it can’t know if what it’s saying is true or not, but because it is being presented as something you can converse with, it feels like there is some deeper cognition that you can interrogate
The more I see & hear, the more I think its all grift.
Ie the crypto bros left their coins for nfts, and now they’ve tanked they’re finding something else to burn the planet down in order to scam suckers.
I don’t think it’s all grift - there are absolutely places where LLMs are the best tech out there, but it’s probably not going to take everyone’s jobs any time soon (at least not on merit - in sure there are plenty of places that’d accept a 50% drop in quality for a 90% drop in price)
I’ve seen a pretty compelling case study of a company using an LLM as a “tier zero” support tech - instead of getting a tier 1 tech to classify a case, decide if they had the tools to address the issue or if it needs to go to tier 2, work out if it was an instance of a known issue etc before they actually start working on the problem, give the LLM some examples and get it to do the triage so the humans can do the more complicated stuff. It does about as well as a human, for a fraction of the price.
I’d have to see that in action before I pass judgement but given LLMs predilection for hallucination and the vagaries of how humans report tech faults I would be surprised if it was significantly more accurate or effective than a human. After all if its working out if there’s a known issue then essentially its not much beyond a script at that point and in that case do you want to trade the unpredictability of what an LLM might recommend vs something (human or otherwise) that will follow the script?
Even if an LLM were an effective level 0 helpdesk it would still need to overcome the user’s cultural expectation (in many places) that they can pick up the phone and speak to somebody about their problem. Having done that job a long long time ago, diagnosing tech problems for people who don’t understand tech can be a fairly complex process. You have to work through their lack of understanding, lack of technical language. You sometimes have to pick up on cues in their hesitations, frustrated tone of voice etc.
I’m sure an LLM could synthesis that experience 80% of the time, but depending on the tech you’re dealing with you could be missing some pretty major stuff in the 20%, especially if an LLM gives bad instructions, or closes without raising it etc. So you then need to pay someone to monitor the LLM and watch what its doing - at which point you’ve hired your level 1 tech again anyway.
Its probably on the same scale as CVs, which all those people struggling with Excel probably wrote “Good Excel skills” on their resume.
I don’t know about my terrible CV but I’ve always claimed I forgot everything about excel - man, I hate the thing passionately.
Having said that I fully support embellishing on ones CV 'cos the role descriptions are word salad or what could best be considered ‘creative writing’
I also hate Excel, but it’s more of a love/hate relationship.
I’ll always remember helping someone with excel and having to explain “yes, I know you’ve gone through some nice GUI menu and come to this field asking you for the date, but you gotta write in 42654.523, so it’s easier if you just ask me every time instead of me trying to teach you why it’s this way.”
I also love/hate excel. It is great for a lot of simple jobs where writing code would take to much time, it is terrible because you can’t audit your code* easily or at all. You get these hideously complex sheets referencing who knows what with no documentation…
- By “code” I mean shitty formula hidden in cells.
That’s what excel is; code for people who don’t know they’re writing code - and its clearly a bad way of doing most of the things people do with it.
But on the flipside you have to give it props for getting people a foot into programming, even if they don’t realise that’s what they’re doing (and folks who use actual languages and lines of text to achieve the same thing don’t accept it for what it kinda is).
I think you could make an argument that Excel is the world’s most used/successful IDE ;)
Haha oh boy, if there’s one eternal rule with complex excel sheets it’s that no one other than the person who made it will ever truely understand it 😆. But you can write VBA functions and throw them in the formulas to make them a bit tidier.
You can…not everybody can.
Not after 3 months - that’s about when I’ve lost everything mentally.
Maybe it was all that acid in the 60s… (Hint: I’m nowhere near 80)
“AI” for health is already known to be very problematic, and nobody wants to see a 10 years down the track commission of inquiry about why some women were not diagnosed correctly from their mammograms.
The chatbots Judith is talking about for tutoring children regularly hallucinate and come up with such stupid things as cooking recipes for petrol spaghetti and other reckless trash. Sounds like a fast way to destroy the education of a bunch of children, but all the rich kids will still be in their private schools with low pupil numbers and enjoying private tutors so why would Judith care.
Ever heard a politician talk about technology and thought “they sound like they really understand this”?
Of course the party that’s “good for business” would be sucked in by that latest CxO craze of “AI everything”.
They can’t but help fall over themselves chasing a technology that promises fewer employees.
There’s a follow up article on RNZ I think talking about this a bit more. One of the ideas someone had is that it could be great in health because they could use AI chatbots to talk to patients in their own language.
Which would be a great service for sure; but like, translation tools already exist and are likely to be better than anything branded as “AI” comes up with for a long time and there’s always like translation services with humans we could just pay to do it without burning the planet.