Marc Daalder 😷 Wear a Mask: You've heard of ChatGPT, what about ChatGOVT? The New Zealand Government could build its own, bespoke AI large language model, Digitising Government Minister Judith Collins told me last week. newsroom.co.nz/2024/07/01/t...

You've heard of ChatGPT, what about ChatGOVT? The New Zealand Government could build its own, bespoke AI large language model, Digitising Government Minister Judith Collins told me last week. newsroom.co.nz/2024/07/01/t...

Govt could develop its own AI, Collins suggestsnewsroom.co.nz A bespoke artificial intelligence for government work would help address concerns about privacy and the misuse of sensitive data by tech companies.

What she means is the pay PWC $10m to do a scoping and engagement exercise for a “one size fits all” approach then drop $100m to develop something so broken and old that they need to pay Infosys $20m a year to run.

It's so clear that Collins, like most people, has no idea how "AI" works. She's fallen for the hijacked sci-fi name, and dreams up magical things that ML and generative models simply cannot do.

Absolutely clueless. She is jumping on a bandwagon that won’t deliver and the name for this is magical thinking.

Compared to the last big hype cycle (blockchain) what makes this one more insidious is that it does have genuinely useful applications across broad range of industries and anyone can try it out. But without understanding even a little of how it works, it's magic so any crazy idea sounds plausible.

💯

Luxon already sounds like predictive text

Lots of hallucinations for sure. I hope he watches some of his own pronouncements. Word salads only good for bingo games

About halfway through I was sort of hoping for some analysis or critique of Collins’s confused claims, e.g. the biggest problem with LLMs is false information, not privacy breaches. Perhaps that will come.

She’s all aboard the hype train created by self-appointed AI experts.

I'm just going to leave this here for anyone interested: link.springer.com/article/10.1...

ChatGPT is bullshit - Ethics and Information Technologylink.springer.com Recently, there has been considerable interest in large language models: machine learning systems which produce human-like text and dialogue. Applications of these systems have been plagued by persist...

NatGPT already spouts gibberish.

This is someone who had to have AI explained to her in simple terms by officials only a few months ago.

TBF, AI couldn't govern worse that the Kaos Klown Koalition....

All Chatgpt does is train people to be dependent on it. Any kid using that as a tutor will just outsource their thinking to it. They won't learn a damn thing.

Never mind the article. What was going on here???

This sounds like a very expensive way to get wrong answers.

Did she say what (probably stolen) data she would train her stochastic parrot aka Large Language Model aka "AI" on? Not objecting to idea of NZ Govt innovating but I would want it to be done ethically.

Judith has a flexible relationship with ethics so don't bet on it.

Christ on a bike

Sounds like magic

What about the energy consumption? www.bbc.com/news/technol...

Warning AI industry could use as much energy as the Netherlandswww.bbc.com A new study looks at the environmental impact of artificial intelligence if current trends continue.

Post