TRASCRIZIONE 🔗
1
00:00:00,094 --> 00:00:03,178
The Nelli Stevenson show is recorded on Wurundjeri Country.
2
00:00:16,221 --> 00:00:20,773
Australia is about to literally explode with data centers that nobody asked for.
3
00:00:20,773 --> 00:00:25,234
They are so problematic that every guest on my show so far this season has brought them
up.
4
00:00:25,234 --> 00:00:28,236
And I wanted to get to the bottom of the data center boom.
5
00:00:28,236 --> 00:00:38,049
A new Greenpeace report released recently calls them energy vampires and outlines the many
frightening ways in which they're sucking our country's energy grid dry while also locking
6
00:00:38,049 --> 00:00:40,560
in polluting gas for decades to come.
7
00:00:40,560 --> 00:00:41,419
Today's guest
8
00:00:41,428 --> 00:00:45,430
is Greenpeace Australia Pacific's head of climate and energy, Joe Rafalowicz
9
00:00:45,430 --> 00:00:48,578
Joe, thank you so much for coming on the show today to chat data centers.
10
00:00:48,578 --> 00:00:49,643
Thank you for having me, Nelli.
11
00:00:49,643 --> 00:00:50,407
It's good to be here.
12
00:00:50,407 --> 00:00:56,702
Look, I'm really keen to chat to you about this report that you've just recently put out
called Energy Vampires.
13
00:00:56,702 --> 00:01:06,967
One of the first things that really jumped out to me was that the largest proposed data
center in Australia is the equivalent in energy use of two million homes.
14
00:01:06,967 --> 00:01:11,570
Just how big are these data centres and who's actually asking for them?
15
00:01:11,570 --> 00:01:22,627
Yeah, so the one of the largest data centers, because they keep getting bigger every week,
it seems, is the Mamre Road proposal in Sydney, and that's one point two gigawatts,
16
00:01:22,627 --> 00:01:26,169
because they measure the size of these data centers based on their energy use.
17
00:01:26,169 --> 00:01:33,441
So one point two gigawatts of power, the emissions from that would be equivalent to five
hundred thousand cars on the road in Sydney.
18
00:01:33,441 --> 00:01:38,276
or all of the domestic flights taken in Australia or New South Wales over a year.
19
00:01:38,276 --> 00:01:41,159
So these are mega scale facilities.
20
00:01:41,159 --> 00:01:48,425
And to put it in context, for a long time data centers in Australia and the world have
been a fraction of that size.
21
00:01:48,425 --> 00:01:57,164
I've found one source that estimated in 2005 Australia had 40 megawatts total of data
center capacity in the whole country.
22
00:01:57,164 --> 00:02:00,737
and now we're talking about 1.2 gigawatts for a single facility.
23
00:02:00,737 --> 00:02:02,798
So that's obviously like many multiples.
24
00:02:02,798 --> 00:02:07,172
And there are now tens, twenties of facilities of that size.
25
00:02:07,172 --> 00:02:16,120
Some most of them are in the kind of 200 megawatt to 400 megawatt range, but there are
some of these what they call hyperscale data centers being proposed now.
26
00:02:16,120 --> 00:02:23,687
and so the reason that we are seeing these much bigger data centers and many more of them
is the growth of
27
00:02:23,687 --> 00:02:28,852
AI as an application, creating all this additional demand on computing power.
28
00:02:28,852 --> 00:02:38,168
Because when you put things in the cloud, like you might save something on Google Drive,
or you know, maybe you're keeping your business customer data in a database with
29
00:02:38,168 --> 00:02:44,562
Salesforce, that is some amount of data, and that's the traditional way that data centers
were utilized.
30
00:02:44,562 --> 00:02:48,164
But when you put a query into Chat GPT,
31
00:02:48,391 --> 00:02:55,834
To use an AI system to answer that query is so much more energy intensive that it requires
much more computing power.
32
00:02:55,834 --> 00:02:59,572
And so that's why we're seeing this frenzied build out of data centers now.
33
00:02:59,572 --> 00:03:11,454
So Joe, you mentioned Chat GPT, and one of the things that I'm really aware of in the line
of work that I do is the way corporations tend to blame their pollution on individual
34
00:03:11,454 --> 00:03:12,034
demand.
35
00:03:12,034 --> 00:03:22,343
And something I'm curious about is how much of this demand for AI data centers is actually
used by, a mum trying to figure out how to get grass stains out of her kids' football
36
00:03:22,343 --> 00:03:29,028
shorts on Chat GPT versus, Google deciding that I need AI in my inbox.
37
00:03:29,028 --> 00:03:36,504
Like what's the sort of percentage of personal use versus all of these corporations just
suddenly deciding AI has to be everywhere?
38
00:03:36,504 --> 00:03:37,436
How does it stack up?
39
00:03:37,436 --> 00:03:41,108
It's a harder question than it sounds to answer.
40
00:03:41,108 --> 00:03:50,834
The Data Center lobby put out a report uh a year or so ago called Data Centers Enabling In
Infrastructure for Our Digital Lives, something like that.
41
00:03:50,834 --> 00:04:01,800
And it had a little story about a woman who wakes up in the morning and she Googles
something that she puts the Netflix on, she uses ABC Listen and listed all these uses of
42
00:04:01,800 --> 00:04:03,389
digital services.
43
00:04:03,389 --> 00:04:07,051
But none of those services would require a hyperscale data center.
44
00:04:07,051 --> 00:04:12,173
In fact, they already exist and they're being powered based on our existing data center
footprint.
45
00:04:12,173 --> 00:04:21,296
So that to me was the first red flag about the the data center lobby not really playing a
straight bat on what's happening in these data centers.
46
00:04:21,296 --> 00:04:32,881
And since then I've seen analysis from McKenzie, the consulting firm, that estimates that
I think it's 70% of data center use will be for AI applications by 2030.
47
00:04:32,881 --> 00:04:39,337
And already the main anchor clients for these big data centers are the cloud service
providers.
48
Load full transcript
00:04:39,337 --> 00:04:45,165
So that's like Microsoft, Google, Meta, and those companies are using their service for
AI.
49
00:04:45,165 --> 00:04:47,087
So a lot of it is being used for AI.
50
00:04:47,087 --> 00:04:52,453
And then on the customer end, who's actually wanting to use AI?
51
00:04:52,453 --> 00:04:59,167
I mean, that's an interesting question because it has been embedded across many
applications that people use.
52
00:04:59,167 --> 00:04:59,368
And
53
00:04:59,368 --> 00:05:03,331
Since writing this research report, I've actually de-googled my personal life.
54
00:05:03,331 --> 00:05:06,414
It was something that I had been meaning to do for a while.
55
00:05:06,414 --> 00:05:10,458
And like I just vaguely felt uncomfortable about it from a privacy perspective.
56
00:05:10,458 --> 00:05:20,677
But once I saw the non-optional embedding of AI, I was like, this has actually gone too
far, like, not just from a privacy perspective, but I don't want my just day-to-day use of
57
00:05:20,677 --> 00:05:24,452
digital infrastructure to be causing all this unnecessary energy use.
58
00:05:24,452 --> 00:05:24,902
As for
59
00:05:24,902 --> 00:05:25,143
So
60
00:05:25,143 --> 00:05:26,803
Like where's the demand coming from?
61
00:05:26,803 --> 00:05:33,995
Like I said, it's getting embedded in all these uh businesses and it's getting embedded in
consumer applications.
62
00:05:33,995 --> 00:05:39,018
If you were given a choice, you know, like it's hard to say would you be using AI or not.
63
00:05:39,018 --> 00:05:50,734
one thing I've learned in this degoogling of my life and also trying to adopt better
habits around AI use, I trialed this tool called Green GPT, which is like a 100%
64
00:05:50,734 --> 00:05:52,265
renewables powered tool.
65
00:05:52,265 --> 00:06:00,009
And the first thing I noticed was not that the tool was, obviously inferior or or
different in some way.
66
00:06:00,009 --> 00:06:09,753
It was how much the software forced me to write proper prompts and strongly discouraged me
from doing a conversational thread with the AI.
67
00:06:09,753 --> 00:06:13,774
It would even show a little sliding tool by you're going to use this much energy.
68
00:06:13,774 --> 00:06:20,913
And I realized that the way that the AI companies have promoted this technology, the
design choices that they've made.
69
00:06:20,913 --> 00:06:26,506
Have encouraged a style of use, which is like, you know, ask a question, ask a follow-up
question.
70
00:06:26,506 --> 00:06:31,900
And as you add a new question, the context window for for the LLM gets longer and longer.
71
00:06:31,900 --> 00:06:40,116
That means more computing power, exactly for the large language model, which is what they
call, chat GPT or anthropic as Claude software.
72
00:06:40,246 --> 00:06:47,373
So our habits are disconnected from the energy use of the tools that we're using because
these companies are embedding it everywhere.
73
00:06:47,373 --> 00:06:50,116
They're encouraging us to have these kind of conversations.
74
00:06:50,116 --> 00:06:52,839
But as a technology, it doesn't have to be like that.
75
00:06:52,839 --> 00:06:58,364
And so Green GPT taught me like, okay, you can use AI if you've got a good use case for
it.
76
00:06:58,364 --> 00:07:02,129
Spend a lot of time thinking about how you're going to use it and try to one shot it.
77
00:07:02,129 --> 00:07:03,750
so that was a
78
00:07:03,822 --> 00:07:10,405
a learning for me that I think is relevant to this conversation and that AI is being
pushed everywhere, it's embedded everywhere, and the way that we're being encouraged to
79
00:07:10,405 --> 00:07:13,070
use it is really not very prudent.
80
00:07:13,070 --> 00:07:16,813
And so I think that's partly behind, the high demand for it.
81
00:07:16,813 --> 00:07:23,936
But as they start charging the real compute costs, because it's a bit like Uber, like when
it first started, it was so cheap everyone started using it.
82
00:07:23,936 --> 00:07:28,418
As they start charging the real cost of it, I think we might see use go down.
83
00:07:28,418 --> 00:07:32,114
Especially if the technology doesn't actually get better, like doesn't actually start
84
00:07:32,114 --> 00:07:35,270
generating value 'cause right now I don't think it is.
85
00:07:35,568 --> 00:07:37,032
That is absolutely fascinating.
86
00:07:37,032 --> 00:07:40,583
So tell me, how does one de Google their life?
87
00:07:40,583 --> 00:07:41,895
How did you do that?
88
00:07:42,449 --> 00:07:45,292
Well, it it kind of sounds easy.
89
00:07:45,292 --> 00:07:46,654
All right, change your email address.
90
00:07:46,654 --> 00:07:52,201
But once you start, you realize how embedded it is into your life.
91
00:07:52,201 --> 00:07:58,068
And so I found the hardest thing to get off totally is Google Maps, just because it's very
convenient.
92
00:07:58,068 --> 00:08:00,878
It's my go-to mapping software and Google Photos.
93
00:08:00,878 --> 00:08:03,302
is shit house anyway, like let's be honest.
94
00:08:03,302 --> 00:08:07,128
How many times have have you gone to a completely different location?
95
00:08:07,128 --> 00:08:09,120
Because Apple Maps just sucks.
96
00:08:09,573 --> 00:08:10,575
Yeah, exactly.
97
00:08:10,575 --> 00:08:15,278
these companies control what is in some ways like public infrastructure.
98
00:08:15,278 --> 00:08:22,283
It's like, you know, using a mapping software is kinda like just it's something that most
people I know do every single day.
99
00:08:22,283 --> 00:08:33,691
And if there were even slight tweaks to it, like say, um, actually I saw a an example uh
that someone shared online of this on Instagram where they noticed Google Maps was sending
100
00:08:33,691 --> 00:08:37,065
them down a less efficient walking route in Ireland.
101
00:08:37,065 --> 00:08:40,516
And so they decided to take the what they thought was the more efficient route.
102
00:08:40,516 -->
00:08:42,587
And sure enough, it cut ten minutes off their trip.
103
00:08:42,587 --> 00:08:44,718
And that street was filled with mansions.
104
00:08:44,718 --> 00:08:50,612
And so their theory was that um some very rich tech investors had prevailed on Google to
change their path.
105
00:08:50,612 --> 00:08:58,796
Anyway, the point is that these things that we trust as public infrastructure are
privately owned, and the owners of them can tweak them in all in all kinds of ways.
106
00:08:58,796 --> 00:09:04,283
They could direct traffic just, you know, one percent more down a certain road that we'll
see a certain billboard, for example.
107
00:09:04,283 --> 00:09:07,425
And there's nothing in their terms of service that would stop them from doing that.
108
00:09:07,425 --> 00:09:20,032
Um yeah, and these things are you know common across as more and more of our life is
online, we're kind of seeding more decisions that we don't even see to these companies
109
00:09:20,032 --> 00:09:20,642
basically.
110
00:09:20,642 --> 00:09:25,365
So I had this interest in just trying to reduce my exposure to that.
111
00:09:25,365 --> 00:09:31,366
So I used uh Proton Mail, which is based in Switzerland, so they have these EU laws.
112
00:09:31,366 --> 00:09:33,147
You're not a hundred percent protected.
113
00:09:33,147 --> 00:09:37,059
Like if if there's some kind of subpoena or something, then they're gonna give your stuff
up.
114
00:09:37,059 --> 00:09:40,741
But I just felt like the more I could get away from big tech, the better.
115
00:09:40,741 --> 00:09:47,834
So I found their mail application quite good and I've s I've used their file storage
software as well.
116
00:09:47,834 --> 00:09:54,607
Cause the thing about Google is that a lot of your data is getting fed into the large
language models to train them.
117
00:09:54,607 --> 00:09:56,914
And you can disable that, but
118
00:09:56,914 --> 00:10:05,703
The assumption should be for most people listening that if you're using a Google service,
unless you've checked all of the the options, that your data is in some way feeding the
119
00:10:05,703 --> 00:10:07,114
training of AI models.
120
00:10:07,114 --> 00:10:09,176
And I just had a problem with that.
121
00:10:09,374 --> 00:10:10,424
Yeah, absolutely.
122
00:10:10,424 --> 00:10:19,103
Actually, fun fact when I was running investigations at Greenpeace working with Joe, I
made sure that an investigations email was set up through Proton Mail, purely based on the
123
00:10:19,103 --> 00:10:22,517
fact that it was the most secure of the email platforms I could find.
124
00:10:22,517 --> 00:10:23,308
Yeah, that's it.
125
00:10:23,308 --> 00:10:33,506
I mean, I've thought about, you know, doing organizing against AI and data centers, not
necessarily for Greenpeace, but just in general, and moving it entirely as much as
126
00:10:33,506 --> 00:10:42,065
possible offline and having a mailing list which is a bunch of names written down
on a
piece of paper and then every month you might send them a postcard.
127
00:10:42,065 --> 00:10:50,799
Like I detect in Gen Z and some of the the younger people that are engaging with activism
a kind of burnout from the digital world.
128
00:10:50,799 --> 00:10:59,463
and a desire to move to IRL experiences and I think it would be invigorating to capture
that in a social movement.
129
00:10:59,463 --> 00:10:59,965
But
130
00:10:59,965 --> 00:11:00,816
interesting.
131
00:11:00,816 --> 00:11:08,012
It's interesting that you mentioned the IRL in real life experiences because I just got an
email from 350 Australia.
132
00:11:08,012 --> 00:11:12,665
90 people turned up to their data center town hall last night.
133
00:11:12,665 --> 00:11:22,401
And anyone who's been involved in any sort of organizing or activism knows how fucking
hard it is to get 90 people out of the house on a Wednesday night in the middle of winter.
134
00:11:22,401 --> 00:11:23,093
Like
135
00:11:23,093 --> 00:11:25,894
Ninety people turned up and they did a postcode analysis.
136
00:11:25,894 --> 00:11:27,755
They were from all over Melbourne.
137
00:11:27,755 --> 00:11:29,846
They were from suburban Melbourne.
138
00:11:29,846 --> 00:11:31,276
They were from inner city Melbourne.
139
00:11:31,276 --> 00:11:33,617
They were from perirregional Melbourne.
140
00:11:33,617 --> 00:11:40,300
And, you know, th it just showed that there are a really huge diversity of people quite
concerned about data centers.
141
00:11:40,300 --> 00:11:50,083
And I wonder, yeah, I wonder that maybe this this is sort of the future of of organizing
that it is gonna become increasingly offline and increasingly more back to these, you
142
00:11:50,083 --> 00:11:51,785
know, real world experiences.
143
00:11:51,785 --> 00:12:00,881
Yeah, I mean I think that data centers have brought together a whole bunch of issues that
were bubbling beneath the surface for many people.
144
00:12:00,881 --> 00:12:06,230
You know, one of them is just the sense of a loss of control of our lives, which has led
to things like
145
00:12:06,297 --> 00:12:08,588
leave campaign and the UK taps into that.
146
00:12:08,588 --> 00:12:11,756
The high vote for Pauline Hanson at the moment reflects
147
00:12:11,756 --> 00:12:14,128
This sense of like the world is out of control.
148
00:12:14,128 --> 00:12:20,262
And I think as you as Donald Trump's presidency careens us from one crisis to the next,
that's increasing.
149
00:12:20,262 --> 00:12:30,018
And then big tech, you know, there's opinion polling on how people perceive big tech from
the heights of kind of when Google said, Don't be evil, that's our corporate motto, to
150
00:12:30,018 --> 00:12:32,270
now, it is absolutely tanked.
151
00:12:32,270 --> 00:12:35,340
And the personal disapproval rating, yeah.
152
00:12:35,340 --> 00:12:37,081
Don't be evil.
153
00:12:37,081 --> 00:12:39,642
Like, don't think of an elephant.
154
00:12:40,283 --> 00:12:42,064
Don't think of an elephant, guys.
155
00:12:42,064 --> 00:12:43,235
Don't think of an elephant.
156
00:12:43,235 --> 00:12:52,130
Like when I train people, when I go and do like trainings now and I train people on
framing, I'll say, Nelli is not a criminal, Nelli is not a criminal, Nelli is not a
157
00:12:52,130 --> 00:12:53,250
criminal, what am I?
158
00:12:53,250 --> 00:12:54,921
And everyone goes, a criminal.
159
00:12:54,921 --> 00:12:57,423
Like don't be evil.
160
00:12:57,423 --> 00:12:59,994
Google, fuck, do better, guys.
161
00:13:00,107 --> 00:13:09,655
And so like maybe at the time it was like so tongue in cheek that it was kind of funny or
like people they were kind of hanging a lantern on their problem sort of thing, like, you
162
00:13:09,655 --> 00:13:10,746
think big tech's evil?
163
00:13:10,746 --> 00:13:11,667
Actually we're not.
164
00:13:11,667 --> 00:13:23,095
But now we've come full circle and it's like, no, they're legit evil because they're like
donating to Trump's, you know, inauguration parties and uh Bezos Elon Musk, their personal
165
00:13:23,095 --> 00:13:25,217
disapproval rating is tanking.
166
00:13:25,217 --> 00:13:26,057
So
167
00:13:26,081 --> 00:13:31,270
My analysis of this is that in the golden days of the tech era, we put up with them.
168
00:13:31,270 --> 00:13:33,222
Like, you know, they were okay.
169
00:13:33,222 --> 00:13:37,710
And on the environmental front, they were leaders on climate commitments and emissions
reduction.
170
00:13:37,708 --> 00:13:39,098
Like Musk was a good guy.
171
00:13:39,098 --> 00:13:44,713
Bezos was standing on top of a a wind farm that he'd paid for and built, smashing a bottle
of champagne.
172
00:13:44,713 --> 00:13:48,686
Like I remember when I was at Climate Council putting that up on our socials.
173
00:13:48,686 --> 00:13:50,177
Like, fuck yeah.
174
00:13:50,177 --> 00:13:53,210
These big tech dudes are investing in renewable energy.
175
00:13:53,210 --> 00:13:58,684
Like and and now they're like literally evil supervillains.
176
00:13:59,149 --> 00:14:00,261
Yeah, that's right.
177
00:14:00,261 --> 00:14:03,854
And like you can see it with their climate emissions and reporting.
178
00:14:03,854 --> 00:14:09,319
Like they essentially have given up on their goals that they set for themselves just two
or three years ago.
179
00:14:09,319 --> 00:14:18,448
Google said that first of all they set a goal and then the next year of reporting they
said it's a moonshot to reach, you know, a hundred percent renewables, whatever it was.
180
00:14:18,448 --> 00:14:23,142
And then the year after they're like, We're still off track because the moon's getting
further away.
181
00:14:23,493 --> 00:14:29,072
And by that they meant because we have to develop AI, our emissions are gonna go up and
that's non negotiable.
182
00:14:29,072 --> 00:14:33,140
So all of these companies are throwing out their environmental commitments to an extent.
183
00:14:33,140 --> 00:14:40,475
I know wonder how much of that because they were also, you know, trying to do the right
thing on things like diversity, equity and inclusion, right?
184
00:14:40,475 --> 00:14:49,258
so I wonder if that trajectory sort of maps against each other, like trying to do the
right thing, trying to be good corporate social citizens, Trump gets back again.
185
00:14:49,258 --> 00:14:50,828
Fuck it all, right?
186
00:14:50,828 --> 00:14:52,999
Like let's just throw it all out the window.
187
00:14:53,391 --> 00:15:02,107
Yeah, and I think certainly the second term of Trump was a bit of an inflection point
because the first term of Trump, there was actually a joint statement from all these tech
188
00:15:02,107 --> 00:15:02,667
companies.
189
00:15:02,667 --> 00:15:09,792
It was it was called like we're all in it together or something, reiterating the need for
the energy transition to clean energy.
190
00:15:09,792 --> 00:15:13,414
And the second time around it was you didn't hear anything from them.
191
00:15:13,414 --> 00:15:20,080
So they are, you know, reading the zeitgeist and they personally seem to have become
radicalized and been right wing.
192
00:15:20,080 --> 00:15:22,661
But to get back to the data centers thing.
193
00:15:22,757 --> 00:15:31,221
I think people in Australia and around the world and even in the US kind of just put up
with how weird Silicon Valley was because they were off doing their own thing and every
194
00:15:31,221 --> 00:15:39,005
now and then we would get a new iPhone or we got Facebook and it was like, okay, well,
they can believe whatever they want, but now they have created this technology that is
195
00:15:39,005 --> 00:15:47,929
sucking in so much capital and is threatening by their own, you know, ambitions for the
technology to replace jobs at a massive scale.
196
00:15:47,929 --> 00:15:51,322
And suddenly whether they have our interests
197
00:15:51,322 --> 00:15:56,944
at heart or whether they are aligned with us as humans starts to become very relevant to
our lives.
198
00:15:56,944 --> 00:16:01,825
And so these data centers, you know, maybe they're not the biggest water user in a
location.
199
00:16:01,825 --> 00:16:04,045
Maybe that's more like an abattoir or something.
200
00:16:04,045 --> 00:16:07,646
Maybe they're not the biggest electricity user, although some of them are.
201
00:16:07,646 --> 00:16:11,427
The point is that people are like, what is the point of this?
202
00:16:11,427 --> 00:16:13,327
It's just a big warehouse.
203
00:16:13,327 --> 00:16:17,808
We put water in, we put electricity in and and Bezos gets richer.
204
00:16:17,808 --> 00:16:20,380
And what the fuck changed in our actual lives?
205
00:16:20,380 --> 00:16:30,630
So I think it has brought together a lot of like simmering insecurity, and difficult
feelings about the state of the world and make them made them concrete.
206
00:16:30,871 --> 00:16:33,772
because that's another reflection I had like around their energy use.
207
00:16:33,772 --> 00:16:39,978
It's like we have thought of digital things as being in the cloud, which is out there
somewhere.
208
00:16:39,978 --> 00:16:41,680
It's uh ephemeral.
209
00:16:41,680 --> 00:16:47,359
But data centers are now the concrete, tangible reality of.
210
00:16:47,359 --> 00:16:48,710
what the cloud is.
211
00:16:48,710 --> 00:16:52,811
That it there has to be a computer somewhere to do this stuff.
212
00:16:52,811 --> 00:16:57,114
So I think we're seeing a bit of a reaction against that, which crosses the political
spectrum.
213
00:16:57,114 --> 00:16:59,955
there was an anti-data centers rally organized in Melbourne.
214
00:16:59,955 --> 00:17:01,806
I think it was in the west of Melbourne.
215
00:17:01,806 --> 00:17:10,219
And I saw that it was like a bunch of speakers pulled out of it, of green speakers pulled
out of it, because the guy that was organizing it was like racist or had done something
216
00:17:10,219 --> 00:17:12,540
that was a bit like exactly.
217
00:17:12,540 --> 00:17:16,149
But you know, I am not one of these people who totally
218
00:17:16,149 --> 00:17:17,590
dismisses that movement.
219
00:17:17,590 --> 00:17:19,881
Like I was in Melbourne during the lockdown.
220
00:17:19,881 --> 00:17:23,622
Of course I dismiss many of the people and the beliefs of the movement.
221
00:17:23,622 --> 00:17:32,847
But the point is that that audience is a group of people who feel very low level levels of
trust in corporations and government.
222
00:17:32,847 --> 00:17:42,591
And I actually think that the defining metric of the future in terms of politics is how
much do you trust the powers that be to look after us.
223
00:17:42,591 --> 00:17:43,083
And
224
00:17:43,083 --> 00:17:48,926
For from a left-wing perspective, we don't trust corporations and we some of us don't
trust governments very much.
225
00:17:48,926 --> 00:17:50,507
It depends on the government.
226
00:17:50,507 --> 00:17:59,331
And so there was some really interesting messaging research that came out of the UK in the
last few weeks that I think the Minderoo Foundation funded, which looked at attitudes
227
00:17:59,331 --> 00:18:00,491
towards AI.
228
00:18:00,491 --> 00:18:07,955
And they found the big determinant was do you trust that governments or companies will do
the right thing?
229
00:18:07,955 --> 00:18:09,183
And if you don't,
230
00:18:09,183 --> 00:18:13,807
trust at a fundamental level, then you're much more in favor of really harsh regulations.
231
00:18:13,807 --> 00:18:18,610
And like it was something like 60% of the population were in favor of regulation in
general.
232
00:18:18,610 --> 00:18:25,075
And of that 60%, I think 40% were in this low trust category or maybe even higher.
233
00:18:25,115 --> 00:18:34,917
And so I bring that up because sovereign citizens, greenies, NIMBY's all of these people
in their gut feel like, what the hell?
234
00:18:34,917
--> 00:18:36,220
when they see a data center.
235
00:18:36,220 --> 00:18:39,376
So it really crosses a whole bunch of political beliefs and values.
236
00:18:39,376 --> 00:18:43,054
It's not a simple left versus right or clean energy versus fossil fuels.
237
00:18:43,054 --> 00:18:44,727
It's touching on so much.
238
00:18:44,772 --> 00:18:52,288
there's this theory called the horseshoe theory, where that often diametrically opposed
political movements may come to the same policy decisions.
239
00:18:52,288 --> 00:19:00,875
And as problematic as that speech was that Pauline Hanson gave to the press club, she is
calling for greater regulation of data centers.
240
00:19:00,875 --> 00:19:10,020
And that is something that, you know, that's an example of that horseshoe where you've got
people on the extreme right and people on the left going, no.
241
00:19:10,061 --> 00:19:12,412
We don't trust that you guys are gonna do the right thing here.
242
00:19:12,412 --> 00:19:21,668
And actually I noticed one of the things in your report said that in places like Texas,
you've got these massive, oil barons coming together with, the greenies and stuff like
243
00:19:21,668 --> 00:19:24,449
that to fight data centers together.
244
00:19:24,449 --> 00:19:34,095
And I wonder, you know, do you think that this is something that as Australia becomes more
politically divided, do you think that this is something that perhaps could be a way to
245
00:19:34,095 --> 00:19:35,576
bring people back together?
246
00:19:35,576 --> 00:19:37,198
United against data centers.
247
00:19:37,198 --> 00:19:40,731
uh It's possible, and I think it will.
248
00:19:40,731 --> 00:19:43,764
I think the challenge will fall on the Labor Party.
249
00:19:43,764 --> 00:19:44,565
Actually, you know what?
250
00:19:44,565 --> 00:19:50,030
It's a challenge for the Liberals as well, because I reckon this is going to be a real
problem in regional areas, especially.
251
00:19:50,030 --> 00:19:58,880
um I think that the growth of data centers has highlighted the disconnect between elites
and everybody else.
252
00:19:58,880 --> 00:20:00,401
And that's why it's
253
00:20:00,401 --> 00:20:10,109
It's and by elites I include many environmental NGOs and many civil society organizations
because actually I think there's a real reticence to get involved in the fight against big
254
00:20:10,109 --> 00:20:14,072
tech because some of the the funding that flows into the sector is from big tech.
255
00:20:14,072 --> 00:20:16,554
We don't want to be seen as anti-everything.
256
00:20:16,554 --> 00:20:27,493
Um in some ways it feels like we're sort of fighting yesterday's battles, you know,
whereas the world has changed, big tech is aligned with fascism, these technologies have
257
00:20:27,493 --> 00:20:29,446
acquired a sinister.
258
00:20:29,446 --> 00:20:34,700
power to do with surveillance and data processing and AI.
259
00:20:34,700 --> 00:20:42,542
And I think that many civil society organizations are still thinking through things from
10 years ago through that frame.
260
00:20:42,542 --> 00:20:51,086
So where do people go when there is not a formal, inclusive opposition movement that they
can join?
261
00:20:51,086 --> 00:20:55,377
Often they get scooped up by right-wing populists like Pauline Hanson and that
262
00:20:55,497 --> 00:21:00,679
dis-ease with what's going on gets channeled in really unproductive ways.
263
00:21:00,679 --> 00:21:11,693
And so that's also how I see, you know, the cookers joining the fight against data centers
is like we have to give people a meaningful way of joining a movement against data centers
264
00:21:11,693 --> 00:21:13,854
or it will come out in in weird ways.
265
00:21:19,076 --> 00:21:28,572
one of the things that you mentioned in your report was that the fossil fuel corporations
are quietly joining the data center lobby group as members, and that there are early signs
266
00:21:28,572 --> 00:21:35,867
of a data center fueled gas boom in Australia, which will come with massive nationally
significant climate costs.
267
00:21:35,867 --> 00:21:43,417
What can you tell our listeners about the relationship between the gas industry and the
data center industry?
268
00:21:43,508 --> 00:21:55,028
Well, the the first thing to say is that in the US, which is kind of ground zero for the
data center boom, there has been this explosion in behind the meter gas use.
269
00:21:55,028 --> 00:22:04,175
And what that means is that when they build these massive data centers, it's quite hard to
connect it to the grid, even in the US, you know, just from a governance perspective.
270
00:22:04,175 --> 00:22:07,218
But also there just might not be enough power in the system.
271
00:22:07,218 --> 00:22:10,240
But these companies will not take no for an answer.
272
00:22:10,240 --> 00:22:11,079
They're in a
273
00:22:11,079 --> 00:22:16,191
a war against each other a race and they have unlimited money for some reason.
274
00:22:16,191 --> 00:22:18,382
So they build their own gas power.
275
00:22:18,382 --> 00:22:23,773
And there is a shortage of even gas turbines of the type that you would have in a gas
power plant.
276
00:22:23,773 --> 00:22:30,376
So they're using gas engines, like what's called a reciprocating engine, like an engine in
your car but powered by gas.
277
00:22:30,376 --> 00:22:32,677
And so that is enormously polluted.
278
00:22:32,677 --> 00:22:34,257
Yeah, exactly at Bistons.
279
00:22:34,257 --> 00:22:36,298
And so that's enormously inefficient.
280
00:22:36,298 --> 00:22:44,953
And it also produces a lot of air pollution, like not just carbon pollution that causes
global warming, but also particulate matter and nitrogen in particular.
281
00:22:44,953 --> 00:22:53,338
And so Meta has one facility that has 700 of these engines, like pumping away using gas
that they're piping in, you know, just for that facility.
282
00:22:53,338 --> 00:23:00,071
Microsoft did a
joint deal with Chevron to build just a totally gas-powered data center.
283
00:23:00,197 --> 00:23:09,059
Elon Musk, X AI is probably the worst offender because he was using, you know, gas
turbines, but he couldn't even wait for the local EPA to sign it off.
284
00:23:09,059 --> 00:23:11,320
He just brought them on site and started powering them.
285
00:23:11,320 --> 00:23:24,804
And it took a drone flyover from an NGO to catch footage, like heat camera footage of the
uh heat coming off the turbines to well, it is illegal and and he's been taken to court.
286
00:23:24,804 --> 00:23:25,274
Yeah.
287
00:23:25,274 --> 00:23:29,661
And then I think Trump has like pardoned well, Trump has pardoned him because they have
288
00:23:29,661 --> 00:23:31,182
you know, Trump in their corner.
289
00:23:31,182 --> 00:23:34,323
So that is like criminal conduct.
290
00:23:34,323 --> 00:23:42,045
And so XAI is the one operating this data center that's powered by gas, but they sell the
computing power to other companies.
291
00:23:42,045 --> 00:23:46,688
Anthropic just signed a computing deal with XAI.
292
00:23:46,688 --> 00:23:51,759
So there was this quit chat GPT movement that was kind of implying oh move to Claude.
293
00:23:51,759 --> 00:23:57,822
Well, Claude is using 100% illegal gas-fired power plants to run their compute too.
294
00:23:57,822 --> 00:24:00,604
So none of these companies have clean hands.
295
00:24:00,604 --> 00:24:06,927
And when the gas companies caught wind of this, this new emerging demand, they were like,
beauty, awesome.
296
00:24:06,927 --> 00:24:13,380
Not only do we have new demand, we have a socially legitimate reason for gas to continue
to be used.
297
00:24:13,380 --> 00:24:14,581
Because this was their problem.
298
00:24:14,581 --> 00:24:17,313
Gas use has been in decline in Australia.
299
00:24:17,313 --> 00:24:19,134
It's at its lowest level in 20 years.
300
00:24:19,134 --> 00:24:20,288
It's expensive.
301
00:24:20,288 --> 00:24:20,768
polluting.
302
00:24:20,768 --> 00:24:25,063
Yeah, it's actually just not a great energy source if you have other opportunities.
303
00:24:25,063 --> 00:24:29,026
So the gas industry saw this amazing opportunity.
304
00:24:29,026 --> 00:24:34,503
And now especially gas turbine manufacturers are really a big part of the data center
movement.
305
00:24:34,503 --> 00:24:42,310
And if you look at like data center trade magazines, one in every two articles is about
like how can you power your data centers with gas.
306
00:24:42,310 --> 00:24:45,605
So they are, you know, making big bucks from this.
307
00:24:45,605 --> 00:24:54,989
And then the gas companies themselves, especially the fossil fuel drilling and extraction
companies in Australia, well, it's a new thing to put in our investor statements.
308
00:24:54,989 --> 00:25:00,041
So they're talking about the demand for gas is going to mean more LNG is required in Asia.
309
00:25:00,041 --> 00:25:01,712
That's what Woodside's talking about.
310
00:25:01,712 --> 00:25:08,114
Beach Energy, an Australian only producer, is saying Australia's going to need more gas
because of data center demand.
311
00:25:09,000 --> 00:25:19,108
Even Australia's climate change authority, when setting Australia's pollution reduction
targets, set a higher ceiling, and they pointed to the new emissions from data centers as
312
00:25:19,108 --> 00:25:20,559
being related to that.
313
00:25:20,559 --> 00:25:31,309
So the data center lobby is planning for more gas to be used, and the gas companies
themselves are banking on data centers as a great new customer.
314
00:25:31,309 --> 00:25:37,910
So this presents a huge challenge for pollution, but it also raises the question: well,
whose interest is that?
315
00:25:37,910 --> 00:25:38,450
in.
316
00:25:38,450 --> 00:25:40,571
You know, it's gonna make gas more expensive for everyone.
317
00:25:40,571 --> 00:25:42,312
It's gonna be more mean more pollution.
318
00:25:42,312 --> 00:25:45,334
It's gonna mean more gas fields opening up.
319
00:25:45,334 --> 00:25:52,858
Like they're gonna drill more gas, more seismic blasting, more pollution of groundwater to
power data centers for what?
320
00:25:52,858 --> 00:25:58,421
So open AI can, you know, build a new model that potentially makes a whole bunch of people
unemployed.
321
00:25:58,421 --> 00:26:00,402
So this is why there's also this pushback.
322
00:26:00,402 --> 00:26:01,643
It's this trade-off.
323
00:26:01,643 --> 00:26:03,644
These decisions are trade-offs.
324
00:26:03,644 --> 00:26:07,352
And I feel like as Australians we are not participating enough
325
00:26:07,352 --> 00:26:13,390
in the decision making it's getting done to us by big tech companies and by fossil fuel
companies.
326
00:26:13,498 --> 00:26:14,990
It's like something I've always said.
327
00:26:14,990 --> 00:26:18,463
Climate change isn't happening to us, it's being done to us.
328
00:26:18,463 --> 00:26:22,986
People are making these decisions to cause climate change.
329
00:26:22,986 --> 00:26:24,858
And this is another example of that.
330
00:26:24,858 --> 00:26:32,093
These data centers are making decisions to lock in more polluting gas for decades to come.
331
00:26:32,093 --> 00:26:37,098
But what I'm curious about, you've mentioned um the link between AI and job losses.
332
00:26:37,098 --> 00:26:44,865
is this something that the unions are organizing on, or has has there been any talk
between, you know, the environmental movement and the union movement about working
333
00:26:44,865 --> 00:26:45,882
together on this?
334
00:26:45,930 --> 00:26:53,467
The union movement has been participating in the like high level discussions with
government on AI regulation.
335
00:26:53,467 --> 00:27:00,092
I would say the union movement is more interested in the AI end of things rather than the
data center end of things.
336
00:27:00,092 --> 00:27:09,080
Um, because the data center's physical
construction and environmental footprint is kind
of, you know, the the focus of the environment movement.
337
00:27:09,080 --> 00:27:13,633
But the union movement is concerned with are our rights protected at work?
338
00:27:13,633 --> 00:27:17,235
Will we see widespread redundancies from AI use?
339
00:27:17,235 --> 00:27:25,941
But I think a more to me realistic concern around jobs and AI is not actually job losses
en masse.
340
00:27:25,941 --> 00:27:33,246
I know that's what the companies are selling and that's what they hope to achieve and it
might happen, but actually the changing nature of work.
341
00:27:33,246 --> 00:27:42,562
So New South Wales, the parliament passed some Australia leading legislation which had to
do with protecting people's work and their mental health.
342
00:27:42,562 --> 00:27:45,313
when their job is determined by an algorithm.
343
00:27:45,313 --> 00:27:47,164
And I've had some experience of this.
344
00:27:47,164 --> 00:27:50,416
I drove, I moonlighted as an Uber driver at one point.
345
00:27:50,416 --> 00:27:56,639
There is something weird about doing a job where you don't have a boss, you have an
algorithm and there's no one to appeal to.
346
00:27:56,639 --> 00:28:03,122
So this legislation said that if your job is could like determined by a computer program,
you have a kind of right of appeal to a human.
347
00:28:03,122 --> 00:28:08,310
You cannot be in a position where you don't have a a boss, a human in the loop.
348
00:28:08,310 --> 00:28:13,884
And it also created protections for where computer systems create unsustainable workloads.
349
00:28:14,104 --> 00:28:19,029
And this has also been a focus of Amanda Rishworth, the workplace minister.
350
00:28:19,029 --> 00:28:31,918
She has been looking at burnout and psychological harm created by AI tools in the
workplace, which encourage unsustainable working patterns to do with context switching.
351
00:28:32,983 --> 00:28:39,589
I think that reflects something interesting, which is that the AI tools can get more
efficient, but there has to be a human in the loop.
352
00:28:39,589 --> 00:28:46,354
And that human then has to manage 50 AI tools and they're not really like implementing
something from start to finish.
353
00:28:46,354 --> 00:28:49,817
They're checking the outputs of AI agents and they're flipping, flipping, flipping.
354
00:28:49,817 --> 00:28:54,982
And they are seeing evidence that that creates a unique kind of psychosocial harm.
355
00:28:54,982 --> 00:29:01,909
So I would say the unions are concerned with, yeah, unemployment, but also the impacts in
the workplace, protecting just like our health and safety.
356
00:29:01,909 --> 00:29:05,441
From a technology, because at the end of the day, it's just another piece of technology.
357
00:29:05,441 --> 00:29:08,279
And so we need to integrate it into our workplaces properly.
358
00:29:08,279 --> 00:29:12,496
I would say the union movement is not against just like data centers in general.
359
00:29:12,496 --> 00:29:14,407
They're kind of like open to it.
360
00:29:14,407 --> 00:29:20,521
I think there is a really strong narrative from government and from the companies that
this is inevitable.
361
00:29:20,521 --> 00:29:23,293
Uh, they say this is happening.
362
00:29:23,293 --> 00:29:26,166
If it doesn't happen here, they're just gonna build it over there.
363
00:29:26,254 --> 00:29:27,947
I don't think that's true.
364
00:29:27,955 --> 00:29:29,046
Yeah, that's right.
365
00:29:29,046 --> 00:29:31,017
And it's
366
00:29:31,069 --> 00:29:42,255
Disempowering, you know, even it's like uh it's ignoring the sovereign ability of a
country like Australia to decide for ourselves if data centers are a good idea.
367
00:29:42,255 --> 00:29:54,322
And I think that is also what arks people up because it's like who asked us before they
decided to build a one gigawatt data center, or even to put AI in my Gmail, or even to
368
00:29:54,322 --> 00:29:59,405
have my photos that I put on Facebook be used to train an AI.
369
00:29:59,411 --> 00:29:59,741
thing.
370
00:29:59,741 --> 00:30:02,021
I was thinking about this the other day.
371
00:30:02,021 --> 00:30:11,329
I felt like this sadness at the fact that everything humans have ever written, sung,
recorded has now been fed into this software.
372
00:30:11,329 --> 00:30:20,497
I think Elon Musk said that by 2027, so next year, they will have run out of text to give
to AI software to train it.
373
00:30:20,497 --> 00:30:23,619
all of human data has now been put into this.
374
00:30:23,619 --> 00:30:26,452
And it's just kind of sad that like we've sort of
375
00:30:26,452 --> 00:30:28,534
given it all to a private company.
376
00:30:28,534 --> 00:30:33,168
It's like moved from the public domain to the private for their benefit, largely.
377
00:30:38,399 --> 00:30:41,309
This has been the most fascinating conversation, Joe.
378
00:30:41,309 --> 00:30:43,299
I'm so grateful that you've come on the show.
379
00:30:43,299 --> 00:30:47,442
I guess what I would really love to pick your brain about now is
380
00:30:47,442 --> 00:30:49,204
What can we do?
381
00:30:49,204 --> 00:30:54,231
What can we do about this this rush to data centers?
382
00:30:54,231 --> 00:30:59,158
And you know, where does the everyday person come into this?
383
00:30:59,158 --> 00:31:01,541
How can we do something about this?
384
00:31:01,541 --> 00:31:03,784
How can we get politicians to listen?
385
00:31:03,784 --> 00:31:05,457
What is Greenpeace calling for?
386
00:31:05,477 --> 00:31:12,350
So right now there is not a nationally consistent, firm set of rules on data centers.
387
00:31:12,350 --> 00:31:15,381
It is the Wild West for the approvals.
388
00:31:15,381 --> 00:31:16,721
state governments do them.
389
00:31:16,721 --> 00:31:19,412
There's no like limit on water or electricity.
390
00:31:19,412 --> 00:31:20,874
It's just all case by case.
391
00:31:20,874 -->
00:31:25,595
We cannot allow a frenzied build out of data centers like that.
392
00:31:25,595 --> 00:31:32,497
So Greenpeace is calling for a moratorium on approvals until there are rules in place that
properly regulate them.
393
00:31:32,823 --> 00:31:38,938
Actually, I think the federal government is receptive to that and they are moving towards
those kind of rules.
394
00:31:38,938 --> 00:31:45,402
But the data center lobby is one of now the wealthiest, most powerful lobbies in the
world.
395
00:31:45,402 --> 00:31:48,365
The amount of money that they have to play with is enormous.
396
00:31:48,365 --> 00:31:52,848
So we require an equivalent mobilization of people.
397
00:31:52,848 --> 00:32:00,292
But luckily, you don't need to make a submission against every data center being proposed
in your area or even attend a rally, but I would encourage that.
398
00:32:00,292 --> 00:32:05,176
You can simply write to your MP and say, hey, I'm concerned about the data center build
out.
399
00:32:05,176 --> 00:32:08,690
I want mandatory rules in place around renewable energy use.
400
00:32:08,690 --> 00:32:15,695
I want there to be strict transparency requirements around water use and access to water
in in drought prone areas.
401
00:32:15,696 --> 00:32:25,957
It's as simple as that because MPs do not get enough pushback, or they need to get a
critical mass of pushback before they start asking questions of the PM.
402
00:32:25,957 --> 00:32:35,425
So even though the data center boom has been massive and it's all over the news, I
actually think politically it's weaker than it looks because who is their constituency?
403
00:32:35,425 --> 00:32:38,498
You know, like it's just money from overseas.
404
00:32:38,498 --> 00:32:44,973
Uh so I think that as Australians, we have a really good opportunity here to intercede
before it gets out of control.
405
00:32:44,973 --> 00:32:46,345
So that's data centers.
406
00:32:46,345 --> 00:32:49,137
And that's I think the first step and the most tangible step.
407
00:32:49,137 --> 00:32:53,239
When it comes to AI regulation, it's more complicated.
408
00:32:53,239 --> 00:33:00,511
And I think we are just at the start of a journey that Australia and the whole world will
go on around how it gets regulated and used.
409
00:33:00,511 --> 00:33:12,835
I think that on a company by company level and on a government by uh government level, and
by that I mean state and federal governments, we need to be moving to lower energy use
410
00:33:12,835 --> 00:33:15,096
technology as a preference.
411
00:33:15,096 --> 00:33:18,027
There is not unlimited energy in the world.
412
00:33:18,027 --> 00:33:22,480
There's unlimited sun and there's unlimited wind, but it's actually not possible
413
00:33:22,480 --> 00:33:25,152
For us to just keep building more and more projects.
414
00:33:25,152 --> 00:33:28,754
Part of the energy transition is using less electricity.
415
00:33:28,754 --> 00:33:32,115
It's not just make
it all renewables, it's use less.
416
00:33:32,115 --> 00:33:39,219
So we need to get back to that mindset of being like, well, this is a trade-off or in a
resource constrained environment.
417
00:33:39,219 --> 00:33:48,365
So there should be rules, I think, for government departments and for companies that you
don't use a LLM, like a large language model that's owned by one of these companies to do
418
00:33:48,365 --> 00:33:49,844
a simple Google search.
419
00:33:49,844 --> 00:33:52,947
There are other ways of of doing that thing.
420
00:33:52,947 --> 00:34:02,777
So it's a bit of a mindset shift, and I don't have the policy solutions for that, but I
think we need to reframe how we're interacting with AI as a society to bring it back down
421
00:34:02,777 --> 00:34:03,537
to reality.
422
00:34:03,537 --> 00:34:07,802
We can't buy into the big tech company's vision of it.
423
00:34:07,802 --> 00:34:09,354
It's not going to be good for us.
424
00:34:09,354 --> 00:34:13,440
Joe Rafalowicz it's just been an absolute joy having you on the show.
425
00:34:13,440 --> 00:34:16,655
This was such an informative and interesting conversation.
426
00:34:16,655 --> 00:34:23,514
And where can people go to learn more about Greenpeace's report and Greenpeace's
campaigning on data centers and AI?
427
00:34:23,584 --> 00:34:23,844
Sure.
428
00:34:23,844 --> 00:34:29,037
Well, if you go to greenpeace.org.au, you'll find the link to the data centers report
there.
429
00:34:29,037 --> 00:34:30,677
It's called Energy Vampires.
430
00:34:30,677 --> 00:34:36,081
You'll notice I didn't encourage you to Google it because that would be using uh
non-optional embedded AI.
431
00:34:36,081 --> 00:34:41,077
and I think that's a good starting point, Climate Council put out a report on data centers
recently.
432
00:34:41,077 --> 00:34:45,980
350.org is, I think, putting together a campaign around a moratorium on data centers.
433
00:34:45,980 --> 00:34:46,990
And there is
434
00:34:46,990 --> 00:34:52,242
uh state by state, location by location opposition to specific data center sites.
435
00:34:52,242 --> 00:34:56,865
So I'd advise you to connect in with those groups if you want to be active in your local
community.
436
00:34:56,865 --> 00:35:04,958
And as we discussed earlier, if you've only got five minutes, you care about this, but you
don't want to be on another email list, just call your MP and be like, hey, I think data
437
00:35:04,958 --> 00:35:08,110
centers are a worry and I would like you to do something about it.
438
00:35:08,110 --> 00:35:12,342
And for those of you who've never called your MP, it's actually incredibly easy.
439
00:35:12,342 --> 00:35:13,893
You literally call them up.
440
00:35:13,893 --> 00:35:17,274
A receptionist type person, an electorate officer will answer the phone.
441
00:35:17,274 --> 00:35:21,806
You'll say, Hey, I'd like to pass a message to such and such MP about data centers.
442
00:35:21,806 --> 00:35:23,577
They go, Yep, just let me get a pen.
443
00:35:23,577 --> 00:35:25,678
They write down what you have to say.
444
00:35:25,678 --> 00:35:27,518
You don't have to be across the policy.
445
00:35:27,518 --> 00:35:29,359
You don't have to have the solutions.
446
00:35:29,359 --> 00:35:32,260
You just say, Look, I live in your electorate.
447
00:35:32,260 --> 00:35:34,001
I'm concerned about data centers.
448
00:35:34,001 --> 00:35:36,347
I want to see this being taken seriously.
449
00:35:37,172 --> 00:35:39,055
And you just say your piece.
450
00:35:39,055 --> 00:35:40,586
and they go, Great, anything else?
451
00:35:40,586 --> 00:35:41,898
And which suburb do you live in?
452
00:35:41,898 --> 00:35:42,218
Great.
453
00:35:42,218 --> 00:35:42,628
Thank you.
454
00:35:42,628 --> 00:35:43,819
I'll pass that on to so and so.
455
00:35:43,819 --> 00:35:45,481
It's a really easy thing to do.
456
00:35:45,481 --> 00:35:49,136
And once you've done it once, you realize how easy it is and you can just do it again and
again.
457
00:35:49,136 --> 00:35:56,087
Yeah, and I'll just make one other point, which is maybe a silver lining of this AI era
that we're in.
458
00:35:56,087 --> 00:35:59,619
As something is more common, it decreases in value.
459
00:35:59,619 --> 00:36:08,477
So I've noticed like email as a form of communication or campaign tactic over time, over
years, has become less valuable because there's just more of it.
460
00:36:08,477 --> 00:36:14,116
It doesn't cost the person sending it as much as it used to, because there used to be
evidence that this person gave a shit.
461
00:36:14,116 --> 00:36:25,471
So that we're seeing that on social media as more and more AI content's generated, even
photos and video, AI content generation means that the person didn't care enough to or
462
00:36:25,471 --> 00:36:28,212
didn't want to put the time in to make it themselves.
463
00:36:28,212 --> 00:36:35,196
So the flip side of that is that real communication, human to human communication becomes
more valuable.
464
00:36:35,196 --> 00:36:41,258
So anytime you can talk to somebody on the phone instead of sending that email, it's gonna
have a bigger impact.
465
00:36:41,262 --> 00:36:51,827
So I think that there is a silver lining to this whole moment in history, which is I sense
a a backlash which might bring us closer together as people rather than driving us further
466
00:36:51,827 --> 00:36:52,213
apart.
467
00:36:52,468 --> 00:36:53,701
Joe Rafalowicz it's it's been a joy.
468
00:36:53,701 --> 00:36:55,475
Thank you so much for coming on the show.
469
00:36:55,553 --> 00:36:56,240
Thank you, Nelli.
470
00:36:56,240 --> 00:36:57,126
A pleasure.
471
00:36:57,248 --> 00:36:59,149
Thanks for sticking around with us on the show today.
472
00:36:59,149 --> 00:37:02,067
If you enjoyed it, please like, subscribe, share it with a friend.
473
00:37:02,067 --> 00:37:08,278
You can follow us on Instagram @thenellistevensonshow and I'm everywhere else just as
Nelli Stevenson.
474
00:37:08,278 --> 00:37:10,208
Have a great day and we'll see you next time.