Can you briefly explain those three problems?

Yeah, so I included in my talk and didn't really have time to elaborate on a range of other ways beyond proxies, all these pieces of data that kind of relate to sensitive personal characteristics, and historic data, all this problem of existing societal divides is directly replicated in training data sets.

I think those are the two that have historically gotten a lot of attention.

But I also pointed out a few subtler ways that we can see inequality being exacerbated potentially by poorly designed, poorly made AI systems.

And yeah, the first is just bad science, right? So this was particularly, I think, overt in say around 2020 when there was a lot of research into facial recognition.

We started to see grossly overstated claims about what these systems could do.

And some of them were just really poor science or replicated pseudoscientific ideas from the past, a kind of fake science that we've really debunked and thrown in the bin.

What else needs to change?

Do you feel the same?

Yeah, I mean, I think one of the reasons why we're having this big conversation about data centers right now is because AI, and particularly generative AI tools, have exploded into people's lives over a really short period of time.

Like, if you think about how long it took, you know, a wide number of people to pick up the smartphone compared to how many people were immediately using ChatGPT upon its launch, like, that rate of increase of uptake is quite staggering.

And even OpenAI was really shocked with how ChatGPT was received.

They weren't expecting it to be as popular as it was.

And there's been so many technical leaps forward that have been really quite astounding.

But at the same time, I think why people have started to push back against data centers is, like, as these tools have become not only more sophisticated but more ubiquitous, so part of our everyday lives, uh, people have also started to get quite fatigued by AI and the constant, uh, promotion of AI and demonization of AI.

If New Zealand wants to be a responsible player in this market, in this AI infrastructure boom, what would a good data center look like?

And what, what are the major concerns that I guess local people can have about data centers and, uh, yeah, how does that manifest itself?

Yeah, that's a good question.

So there's a range of different concerns.

So there's a set of different concerns when it comes to how data centers can impact local residents' quality of life.

So data centers, uh, release pollution into the air, and they also are very noisy.

So they produce this kind of droning or humming sound, uh, which can affect the sort of quality of life for residents who live nearby.

Uh, so th-those are sort of some immediate issues that if you had a data center next to you, uh, you might be worried about.

Yeah

Yeah, and what, what are the major concerns that, I guess, local people can have about data centers and, uh, yeah, how does that manifest itself?

Yeah, that's a good question.

So there's a range of different concerns.

So there's a set of different concerns when it comes to how data centers can impact local residents' quality of life.

So data centers, uh, release pollution into the air, and they also are very noisy, so they produce this kind of droning or humming sound, uh, which can affect the sort of quality of life for residents who live nearby.

Uh, so th- those are sort of some immediate issues that if you had a data center next to you, uh, you might be worried about.

Uh, but there's also a wider range of environmental issues beyond, of course, the important question of pollution.

Um, how, how well informed do you think people are on these data centers in New Zealand, and do you think enough has been done to bring the public on board?

And what are the major concerns that I guess local people can have about data centers? And yeah, how does that manifest itself?

Yeah, that's a good question.

So there's a range of different concerns.

So there's a set of different concerns when it comes to how data centers can impact local residents' quality of life.

So data centers release pollution into the air and they also are very noisy.

So they produce this kind of droning or humming sound.

which can affect the sort of quality of life for residents who live nearby.

How well informed do you think people are on these data centers in New Zealand? And do you think enough has been done to bring the public on board?

And what are the major concerns that I guess local people can have about data centers? And yeah, how does that manifest itself?

Yeah, that's a good question.

So there's a range of different concerns.

So there's a set of different concerns when it comes to how data centers can impact local residents' quality of life.

So data centers release pollution into the air and they also are very noisy.

So they produce this kind of droning or humming sound.

which can affect the sort of quality of life for residents who live nearby.

How well informed do you think people are on these data centers in New Zealand? And do you think enough has been done to bring the public on board?

Well, what alternative do we have? Because isn't the genie out of the bottle? Once you've designed something as powerful as that, it just simply is as powerful as that.

Yeah, I mean, I think the question here shouldn't be putting the genie back into the bottle.

And I think this is often an argument people make to suggest that people who want a different kind of future for AI make to suggest that to ask for anything different would be Luddite or retrograde.

But I think that there's different ways we could be designing and regulating these models to make them safer for everyone.

you know what's really interesting about this case is that companies like open ai are really keen to avoid or undermine open weight or open source models models that people can easily download and amend themselves to protect their own sort of market position to be the leaders in this space But ironically, despite all their advertising previously that open source is really dangerous, the company Hugging Face actually ended up stopping the cyber attack through the use of a Chinese open weight model.

And so I think what it really shows is that we should be very careful about letting our AI regulation and policy and governance be led by a very narrow hand of actors like OpenAI or Anthropic.

who, you know, ultimately, I think, do just have their best interests at heart.

Would you look at those two things and think this is too dangerous to release publicly or no?

Yeah, right.

Uh, I mean, yeah.

I mean, I think it just comes down to... again, like, I think there's a fundamental irony of saying, like, you know, the power of these AI-enabled tools and products is that, you know, we can perform all these, like, massive tasks at scale.

And this idea of, again, like, a product that can be sold across the world, a product that can be used at scale with the kind of knowledge that this only really works for, like, a very, very narrow base.

I mean, my assumption, to be fair though, is, like, I think that a lot of people who make products that have, you know, these kinds of exclusions or biases aren't necessarily going in being like, "I know my product is really biased and I actually just, like, don't care." Like, I don't think that usually is it.

I think often, to me, it's just this kind of mindset of either, "We just, like, haven't really thought about it." I think that's particularly common with accessibility, which is that often accessibility has to be integrated in from the very beginning of the design process.

It can't be slapped on at the end.

That's when I started to begin to be like, hmm, this feels a little bit more like a smoking gun when you hear, like, you know these ideas are being, like, ex- exchanged and knowing that, like, there's this guy, Curtis Yarvin, that we talked about a few weeks ago who's, like, basically, like, a tech monarchist who has a lot of ideas that people like Elon Musk and Peter Thiel, like, are... really subscribe to and- J.D. Vance.

Yeah.

I mean, yeah, I mean, I think it just comes down to again, like I think there's a fundamental irony of saying like, you know, the power of these AI-enabled tools and products is that, you know, we can perform all these like massive tasks at scale.

And this idea of again, like a product that can be sold across the world, a product that can be used at scale with the kind of knowledge that this only really works for like a very, very narrow base.

I mean, my assumption to be fair though is like I think that a lot of people who make products that have, you know, these kinds of exclusions or biases aren't necessarily going in being like, "I know my product is really biased and I should just, I don't care." Like, I don't think that usually is it.

I think often to me it's just this kind of mindset of either we just like haven't really thought about it.

I think this is particularly common with accessibility, which is that of the accessibility has to be integrated in from the very beginning of the design process.

It can't be slapped on at the end.

Right.