Latest / 6 Ps in AI Pods (AI6P) / 🗣️ AISW #130: Marilyn Arceo, USA-based enterprise solutions & data architect

Transcript

🗣️ AISW #130: Marilyn Arceo, USA-based enterprise solutions & data architect · 6 Ps in AI Pods (AI6P)

• 0:00
  SPEAKER_01: Hi everyone.
• 0:00
  SPEAKER_01: This is Karen Smiley.
• 0:02
  SPEAKER_01: Welcome to 6P's in AI Pods.
• 0:07
  SPEAKER_01: This post is part of our AI 6P interview series on AI software and wetware.
• 0:13
  SPEAKER_01: Our guests share their experiences with using AI and how they feel
• 0:17
  SPEAKER_01: about AI using their data and content.
• 0:21
  SPEAKER_02: I'm delighted to welcome Marilyn Arceo from Los Angeles,
• 0:24
  SPEAKER_02: USA as my guest today on AI software and wetware.
• 0:28
  SPEAKER_02: Marilyn thank you so much for joining me on this interview.
• 0:30
  SPEAKER_02: Please tell us about yourself, who you are and what you do.
• 0:34
  SPEAKER_00: Hi, my name is Marilyn Arceo.
• 0:36
  SPEAKER_00: My background is in enterprise solution and data architecture,
• 0:40
  SPEAKER_00: but I actually started off as a chemistry and physics major and wanted
• 0:43
  SPEAKER_00: a double major in art.
• 0:44
  SPEAKER_00: So completely different background.
• 0:47
  SPEAKER_00: And I work in the aerospace and defense industry now,
• 0:51
  SPEAKER_00: but I really started off my career as a software engineer in fashion all
• 0:56
  SPEAKER_00: the way down to now planes and spacecraft.
• 0:59
  SPEAKER_00: So definitely have come a long way.
• 1:02
  SPEAKER_02: Yeah, chemistry and physics I can see in the background for aerospace,
• 1:05
  SPEAKER_02: the art aspect is interesting and maybe we can talk more about that
• 1:08
  SPEAKER_02: and your experiences with it.
• 1:09
  SPEAKER_00: Of course.
• 1:11
  SPEAKER_02: Tell us a little bit about your local experience with AI and machine learning
• 1:14
  SPEAKER_02: and analytics,
• 1:15
  SPEAKER_02: whether you've used it professionally or personally and whether
• 1:19
  SPEAKER_02: or not you've ever studied it either formally or informally.
• 1:23
  SPEAKER_00: Yeah, so I've had experience leveraging ML and image processing early
• 1:28
  SPEAKER_00: in my academic years,
• 1:29
  SPEAKER_00: and so I was able to transition that skill set to a lot
• 1:32
  SPEAKER_00: of my different roles professionally.
• 1:34
  SPEAKER_00: I actually started off building artificial neural networks
• 1:37
  SPEAKER_00: and particle swarm optimization models for my instrument optimization.
• 1:41
  SPEAKER_00: We had a chemistry lab back at my university,
• 1:45
  SPEAKER_00: so we're just trying to predict the most optimal conditions
• 1:48
  SPEAKER_00: to run our capillary electrophoresis instrument.
• 1:51
  SPEAKER_00: And we just had so many different types of research that we're trying to go fast.
• 1:55
  SPEAKER_00: So I dabbled into building a framework for image recognition and predictive models
• 2:01
  SPEAKER_00: for characterizing chemical environments and determining trends.
• 2:05
  SPEAKER_00: Now with that skill set,
• 2:06
  SPEAKER_00: I was able to transfer all that knowledge to building tools
• 2:09
  SPEAKER_00: and analytical tools ranging from BI dashboarding,
• 2:12
  SPEAKER_00: predictive customer retention and churn,
• 2:14
  SPEAKER_00: data engineering,
• 2:15
  SPEAKER_00: and even things like optimal character recognition and predictive applications
• 2:20
  SPEAKER_00: across all these different industries.
• 2:23
  SPEAKER_02: That sounds like you're using it quite a lot professionally,
• 2:25
  SPEAKER_02: mostly what I would say is traditional machine learning,
• 2:29
  SPEAKER_02: not generative AI, not the bots.
• 2:31
  SPEAKER_02: Do you use AI chatbots in any way that you're using those?
• 2:36
  SPEAKER_00: I'm barely now transitioning using LLMs and MCP agents now and trying
• 2:41
  SPEAKER_00: to figure out how we can use that in some of our processes and automation and
• 2:45
  SPEAKER_00: so it's been interesting just because I come from I would say more of
• 2:49
  SPEAKER_00: a traditional background and so it's kind of been like okay this is a lot powerful
• 2:54
  SPEAKER_00: at your fingertips now and it's completely different from how it was so yeah Yeah,
• 3:00
  SPEAKER_02: maybe we can talk a little bit more about that.
• 3:02
  SPEAKER_02: So professionally, you are making some good uses of it.
• 3:05
  SPEAKER_02: How about in your personal life?
• 3:06
  SPEAKER_02: Do you look to use it?
• 3:08
  SPEAKER_02: Do you avoid it?
• 3:09
  SPEAKER_02: I mean, it's very hard to avoid AI entirely in our lives nowadays.
• 3:13
  SPEAKER_00: It honestly just depends, right?
• 3:15
  SPEAKER_00: If I'm researching a topic, right?
• 3:17
  SPEAKER_00: Because I still like to read a lot of my things if I'm doing any types of research.
• 3:21
  SPEAKER_00: I'm usually using an LLM to do extensive research on a specific topic so
• 3:27
  SPEAKER_00: that I can go read through these PDFs and academic journals or industry papers.
• 3:31
  SPEAKER_00: So that's where I use a lot of my AI.
• 3:33
  SPEAKER_00: Personally, I always feel like it's more of a really beefed up search engine.
• 3:38
  SPEAKER_00: I've dabbled in trying to use Claude's code to kind of test out those capabilities, right?
• 3:44
  SPEAKER_00: But I was a little bit hesitant just giving something access to my file system
• 3:50
  SPEAKER_00: and my GitHub and all that kind of stuff.
• 3:52
  SPEAKER_00: So it's been me trying to work through the security aspect of it.
• 3:56
  SPEAKER_02: Yeah, security and governance are huge things once you start enabling, giving them credentials.
• 4:01
  SPEAKER_00: Yeah, I'll randomly get an email from Claude that wants elevated GitHub access
• 4:09
  SPEAKER_00: and I keep declining it.
• 4:10
  SPEAKER_00: It also gives you a little bit of apprehension.
• 4:12
  SPEAKER_00: Like, that's a little weird.
• 4:15
  SPEAKER_02: And why is it asking for that?
• 4:18
  SPEAKER_02: Yeah, least privilege.
• 4:22
  SPEAKER_02: So you've probably heard some of the recent stories.
• 4:24
  SPEAKER_02: Some people say they're more hyped than reality
• 4:26
  SPEAKER_02: about how these different LLMs have escaped containment and broken out and hacked
• 4:31
  SPEAKER_02: into other systems.
• 4:34
  SPEAKER_02: You may have some opinions on that.
• 4:36
  SPEAKER_00: Yeah, I know.
• 4:37
  SPEAKER_00: I think I recently even saw one where it escaped the sandbox and then it was trying
• 4:41
  SPEAKER_00: to attack another LLM and take it down.
• 4:45
  SPEAKER_00: And I'm like, huh.
• 4:47
  SPEAKER_00: So what are the guardrails, right?
• 4:49
  SPEAKER_00: What are we putting in place?
• 4:50
  SPEAKER_00: And what's the lack of governance that we have in cybersecurity?
• 4:54
  SPEAKER_00: Like, guardrails, are we not taking these measures?
• 4:56
  SPEAKER_00: So it's been quite interesting,
• 4:59
  SPEAKER_00: especially since I work in the industry that I work in,
• 5:01
  SPEAKER_00: where cybersecurity first, right?
• 5:03
  SPEAKER_00: Security first.
• 5:04
  SPEAKER_00: I'm always consumer first and privacy first type of person.
• 5:09
  SPEAKER_00: It's a little weird to see that we haven't contained and made it bulletproof
• 5:16
  SPEAKER_00: and it's somehow escaping.
• 5:17
  SPEAKER_00: It worries me as well,
• 5:18
  SPEAKER_00: especially if you talk about people just put things in there without thinking.
• 5:23
  SPEAKER_00: So it's like all your data is now out there or even there is an incident where
• 5:28
  SPEAKER_00: if you share a Claude chat,
• 5:31
  SPEAKER_00: it gets indexed into Google and you can Google search it.
• 5:35
  SPEAKER_00: And so now your data privacy is completely gone.
• 5:37
  SPEAKER_00: It's been really weird just to see all the type of news.
• 5:42
  SPEAKER_02: Yeah.
• 5:43
  SPEAKER_02: I'd like to hear a story about one specific task that you've used an AI tool for.
• 5:48
  SPEAKER_02: It doesn't have to be a chatbot.
• 5:50
  SPEAKER_02: It could be some other type of AI tool.
• 5:52
  SPEAKER_02: Maybe one of your customer term predictions or something.
• 5:55
  SPEAKER_02: Anything like that where you've used machine learning
• 5:57
  SPEAKER_02: and we all know some things go well, some things don't go well.
• 6:01
  SPEAKER_00: Yeah, I think for me it's almost like market research, right?
• 6:05
  SPEAKER_00: I used it to compare almost like a market research report.
• 6:09
  SPEAKER_00: I was comparing different types of software between three different companies
• 6:13
  SPEAKER_00: and figure out who the target audience is,
• 6:16
  SPEAKER_00: what their capabilities are, and see the trends.
• 6:19
  SPEAKER_00: Just because a lot of the time the software gets bought out by private equity and
• 6:23
  SPEAKER_00: so usually you'll see the trends of like Hey,
• 6:25
  SPEAKER_00: if this specific private equity invests in this particular software
• 6:29
  SPEAKER_00: or buys them out,
• 6:31
  SPEAKER_00: what's the likelihood of them just,
• 6:33
  SPEAKER_00: I hate to use a term, but they're going to Joanne it, right?
• 6:36
  SPEAKER_00: And so what are the alternatives, right?
• 6:38
  SPEAKER_00: Just doing that market research and analysis of alternatives and all that research.
• 6:41
  SPEAKER_00: I've been using LLMs for that just because it's been really hard to compile a lot of
• 6:49
  SPEAKER_00: that data really quickly,
• 6:50
  SPEAKER_00: especially if you have to present it to a customer or you're making
• 6:54
  SPEAKER_00: a business decision on what type of software we should be using, right?
• 6:58
  SPEAKER_00: And what should we invest?
• 6:59
  SPEAKER_00: And so a lot of times leadership and customers will pressure you for a really tight deadline.
• 7:06
  SPEAKER_00: And that's one of the ways I've been able to use that is compile my data,
• 7:09
  SPEAKER_00: but also put the human in the loop to make sure that what is compiling is true.
• 7:13
  SPEAKER_00: Verifying it because that's also something I keep thinking about as well.
• 7:18
  SPEAKER_02: Yeah.
• 7:18
  SPEAKER_02: Have you ever found that something that it summarized or presented to you was not accurate?
• 7:23
  SPEAKER_02: And how did you catch it?
• 7:25
  SPEAKER_02: And this has changed in the past say a year or two.
• 7:27
  SPEAKER_02: Is it getting better or worse?
• 7:29
  SPEAKER_00: I think it got better.
• 7:31
  SPEAKER_00: And it just depends, right?
• 7:32
  SPEAKER_00: It just depends on what models you're using and how you're prompting things.
• 7:36
  SPEAKER_00: It definitely was a lot worse before.
• 7:38
  SPEAKER_00: It was hallucinating and making things up.
• 7:40
  SPEAKER_00: I even had an instance where I prompted something to it and I knew the answer to it.
• 7:47
  SPEAKER_00: It was almost like a test and it just made something up and I obviously rated it down.
• 7:52
  SPEAKER_00: Hey, that's wrong.
• 7:53
  SPEAKER_00: That's incorrect.
• 7:54
  SPEAKER_00: And when you correct the LLM like hey that's wrong it's actually this because
• 7:58
  SPEAKER_00: if you read the document that you just gave me it says this
• 8:01
  SPEAKER_00: so usually whenever I prompt something I always ask it for reference material and
• 8:05
  SPEAKER_00: the links that I use to come up with the answer just because I'll go back
• 8:10
  SPEAKER_00: and check it I don't know maybe it's the millennial in me that we always had
• 8:13
  SPEAKER_00: to cite our sources for all our research papers but I always have to go check
• 8:17
  SPEAKER_00: and see oh this is what you said let me just make sure this is actually real life
• 8:21
  SPEAKER_00: and it's definitely gotten better but I do catch it still hallucinating just
• 8:25
  SPEAKER_00: a little bit and it's a little worrisome so it's always tell people trust
• 8:30
  SPEAKER_00: but verify yeah
• 8:32
  SPEAKER_02: market research is interesting too
• 8:33
  SPEAKER_02: because it may be accessing sources
• 8:36
  SPEAKER_02: that aren't necessarily trustworthy it could be somebody is writing hype
• 8:41
  SPEAKER_02: and they've got good SEO or good GEO I guess they call it now Yeah.
• 8:46
  SPEAKER_02: And that's how it got into cod or perplexity or whatever else you're using.
• 8:50
  SPEAKER_00: Right.
• 8:51
  SPEAKER_00: Yeah.
• 8:51
  SPEAKER_00: I always try to reference things that are typically credible, right?
• 8:55
  SPEAKER_00: I know a lot of people use it in McKinsey and all those really big firms
• 8:59
  SPEAKER_00: or even industry partners
• 9:00
  SPEAKER_00: or industry associations are recommending certain practices.
• 9:04
  SPEAKER_00: So I do always try
• 9:05
  SPEAKER_00: to make sure it's referencing this specific poll versus internet chatter
• 9:11
  SPEAKER_00: and blog posts.
• 9:11
  SPEAKER_00: Because I also did notice that if you reference a blog post with someone's opinion
• 9:15
  SPEAKER_00: and it's not quite true and factual.
• 9:19
  SPEAKER_02: It's not a primary source.
• 9:20
  SPEAKER_02: Yeah.
• 9:21
  SPEAKER_00: Yeah.
• 9:22
  SPEAKER_02: It sounds like you're making good use of AI tools in a lot of ways.
• 9:25
  SPEAKER_02: Are there any things that you would avoid using AI tools for?
• 9:29
  SPEAKER_02: And can you share an example of when you would avoid it and why?
• 9:34
  SPEAKER_00: I have avoided using it for any artwork or illustration.
• 9:38
  SPEAKER_00: Again, I'm an artist and I remember when ChatGPT had generative images
• 9:43
  SPEAKER_00: and it first started coming out and I tried it.
• 9:47
  SPEAKER_00: I'm not going to lie, I had to go figure out what it was doing.
• 9:51
  SPEAKER_00: I had trouble prompting it and getting exactly what I wanted.
• 9:55
  SPEAKER_00: And I'm like, I'm just waiting computational resources at this point.
• 9:59
  SPEAKER_00: And I could have figured out how to do this in Canva or Photoshop or using Procreate
• 10:05
  SPEAKER_00: in a matter of a few minutes, right?
• 10:06
  SPEAKER_00: And so I felt like I'm probably myself of the experience of the creativeness, right?
• 10:11
  SPEAKER_00: And you can tell when something's generated with AI.
• 10:15
  SPEAKER_00: It almost looks a little soulless to me, just a little bit.
• 10:17
  SPEAKER_00: And as an artist myself, I can't, I can't bring myself to rely on that.
• 10:21
  SPEAKER_00: I've even volunteered for a campaign and I've done graphic design and all that kind
• 10:27
  SPEAKER_00: of stuff with friends and website design.
• 10:29
  SPEAKER_00: And when I engage with a person directly,
• 10:32
  SPEAKER_00: like a customer essentially,
• 10:34
  SPEAKER_00: and I'm designing for them, we always have better outcomes.
• 10:36
  SPEAKER_00: We kind of figure out the vision and we get there a lot faster.
• 10:39
  SPEAKER_00: And so then when we're generating something through AI and
• 10:44
  SPEAKER_00: Just keep prompting over and over and over again.
• 10:46
  SPEAKER_00: We're never going to get something that we want.
• 10:48
  SPEAKER_00: So to me, it's like time savings at this point.
• 10:52
  SPEAKER_02: So part of it is because of your art background and you have the skills to do it yourself.
• 10:57
  SPEAKER_02: A lot of people obviously don't,
• 10:58
  SPEAKER_02: and they may feel like it's an easy solution, but as you said, it can feel soulless.
• 11:04
  SPEAKER_02: A lot of people have commented about that.
• 11:05
  SPEAKER_02: So one concern that comes up sometimes,
• 11:07
  SPEAKER_02: especially with regard to art,
• 11:09
  SPEAKER_02: is that they train the tools by basically scraping or stealing work
• 11:16
  SPEAKER_02: from professional artists and that isn't fair and some people pull back
• 11:20
  SPEAKER_02: from it really for those reasons.
• 11:21
  SPEAKER_02: It sounds like you're not happy with the results that you get.
• 11:25
  SPEAKER_00: I would say is a little of both, right?
• 11:27
  SPEAKER_00: Because there's even the argument of not only the creative portion of it,
• 11:31
  SPEAKER_00: but the licensing of the art.
• 11:33
  SPEAKER_00: If you look at something like Adobe,
• 11:34
  SPEAKER_00: which I didn't realize until probably the last couple of years,
• 11:37
  SPEAKER_00: where if you upload your images right onto Adobe Creative Cloud,
• 11:41
  SPEAKER_00: I believe that they use your images to help do that generative AI stuff.
• 11:46
  SPEAKER_00: I've been retouching photos since I was like nine.
• 11:49
  SPEAKER_00: So for me,
• 11:50
  SPEAKER_00: I was like,
• 11:51
  SPEAKER_00: why would I have Adobe do it I'll just do it myself because again I've been doing
• 11:55
  SPEAKER_00: that but I didn't realize that those were like the terms and conditions
• 11:59
  SPEAKER_00: that we were signing up for when that whole generative AI thing came out with Adobe
• 12:04
  SPEAKER_00: and so that was a little bit weird for me just because those are my pictures
• 12:08
  SPEAKER_00: so do you technically own them and put them in a model I was like I didn't agree to
• 12:12
  SPEAKER_00: that And no one's really going to read the T's and C's, right?
• 12:15
  SPEAKER_00: And so that's really weird.
• 12:17
  SPEAKER_00: So I guess for me, it's a both thing.
• 12:18
  SPEAKER_00: There's a creative aspect of it.
• 12:20
  SPEAKER_00: And then you're using our work or somebody else's work.
• 12:23
  SPEAKER_00: And I know that's happening also in music as well, like AI music.
• 12:26
  SPEAKER_00: And so you're using both voices.
• 12:28
  SPEAKER_00: And so at what point, if you're using somebody's voice, you're not paying them either.
• 12:33
  SPEAKER_00: It's not licensed either.
• 12:34
  SPEAKER_00: So you should be paying and compensating these artists no matter what.
• 12:38
  SPEAKER_00: You're using their work for your own gain.
• 12:40
  SPEAKER_00: They have billions of dollars.
• 12:41
  SPEAKER_00: They can afford it.
• 12:43
  SPEAKER_02: Yeah, it definitely changes the economics.
• 12:45
  SPEAKER_02: I think it comes down to a question of fairness.
• 12:47
  SPEAKER_02: A lot of artists, the musicians, it's impacting their livelihoods.
• 12:52
  SPEAKER_02: One aspect is the copyright infringement.
• 12:54
  SPEAKER_02: The other aspect is the unfair competition and using their own work
• 12:58
  SPEAKER_02: to basically compete against them.
• 13:00
  SPEAKER_02: And AI companies are making the profit in the middle and not the original artists
• 13:04
  SPEAKER_02: and not the people who would have gotten some work to do that.
• 13:07
  SPEAKER_02: But then there are also cases where for someone creating a picture to attach to their blog.
• 13:14
  SPEAKER_02: They're not going to pay an artist for that anyway.
• 13:16
  SPEAKER_02: So there's a lot of controversy around that.
• 13:18
  SPEAKER_02: You mentioned compensating people.
• 13:20
  SPEAKER_02: There's a concept called the three C's,
• 13:23
  SPEAKER_02: which came from CIPRI, the Cultural Intellectual Property Rights Initiative.
• 13:26
  SPEAKER_02: And they advocate that for creative rights, they should be able to consent.
• 13:31
  SPEAKER_02: It sounds like you did not consent to what he was doing
• 13:34
  SPEAKER_01: with your pictures.
• 13:35
  SPEAKER_01: You should be credited if your
• 13:37
  SPEAKER_02: work is used.
• 13:37
  SPEAKER_02: I know a lot of musicians have said they're not getting credited
• 13:40
  SPEAKER_02: when their work is used and compensated.
• 13:42
  SPEAKER_02: Compensation shows up in all these lawsuits with books and music and art and everything else.
• 13:48
  SPEAKER_02: What are your thoughts about the three C's?
• 13:50
  SPEAKER_02: It definitely would change the dynamic.
• 13:52
  SPEAKER_00: Yeah, I do believe companies that are using the data and content to train their
• 13:57
  SPEAKER_00: and ML systems need to be transparent,
• 14:00
  SPEAKER_00: obviously where their data comes from,
• 14:02
  SPEAKER_00: but also if they're using it, they should be compensating.
• 14:04
  SPEAKER_00: That's the bottom line, right?
• 14:06
  SPEAKER_00: I'm not a big fan of how companies leverage the data,
• 14:10
  SPEAKER_00: but I also do recognize as the reality that they're going to market
• 14:13
  SPEAKER_00: and they're going to sell products to consumer and they're going to build profiles
• 14:16
  SPEAKER_00: on people, they're going to use whatever it is to get business.
• 14:22
  SPEAKER_00: Yeah, I do think they need to be compensating everyone for that.
• 14:26
  SPEAKER_00: It's just not fair.
• 14:27
  SPEAKER_00: It's kind of like Disney's licensing, right?
• 14:29
  SPEAKER_00: Like whenever you buy a product,
• 14:30
  SPEAKER_00: it's a licensed merchandise,
• 14:32
  SPEAKER_00: it's official product,
• 14:34
  SPEAKER_00: and you pay someone a fee to use Mickey Mouse on your product,
• 14:38
  SPEAKER_00: on your socks or whatever it is.
• 14:40
  SPEAKER_00: They should be doing the same thing.
• 14:41
  SPEAKER_00: In terms of the 3 C's I'm also kind of looking at from
• 14:44
  SPEAKER_00: a data broker situation right people are using anonymized data and they're mining it
• 14:50
  SPEAKER_00: from different channels so yeah again it should be disclosed to the people
• 14:56
  SPEAKER_00: and they should understand how their data is being used and where
• 14:59
  SPEAKER_00: that data is being presented and how it's originating something that I like
• 15:04
  SPEAKER_00: to preach about is data governance and lineage and so I know
• 15:08
  SPEAKER_00: the large companies will typically use systems like Databricks or Snowflake
• 15:12
  SPEAKER_00: and build all these data lakes.
• 15:14
  SPEAKER_00: One of the key things that they can and they probably already do
• 15:17
  SPEAKER_00: but they're probably not telling us is they're tagging their data
• 15:19
  SPEAKER_00: and they're maintaining that traceability on it and where it came from in terms
• 15:23
  SPEAKER_00: of integration and all the way down to all the metrics that they're using
• 15:27
  SPEAKER_00: into their dashboards and their ML models.
• 15:29
  SPEAKER_00: I think ethical AI companies should give the consent and give the credit
• 15:34
  SPEAKER_00: and compensate people at the end of the day If they're going to use it for training,
• 15:38
  SPEAKER_00: looking at the creative world,
• 15:39
  SPEAKER_00: musicians, artists, and so on, they own their original work.
• 15:43
  SPEAKER_00: And so again,
• 15:44
  SPEAKER_00: it goes back to that licensing,
• 15:45
  SPEAKER_00: like Disney licensing, like the Mickey Mouse for a pair of socks, right?
• 15:50
  SPEAKER_00: If they're going to license intellectual property for use of marketing
• 15:53
  SPEAKER_00: and product development across different channels, it should apply here.
• 15:58
  SPEAKER_02: Yeah, it's a very fair point.
• 15:59
  SPEAKER_02: Yeah, the Disney licensing is an analogy I haven't heard before, but it makes total sense.
• 16:04
  SPEAKER_02: So as someone who uses these tools,
• 16:08
  SPEAKER_02: do you feel like any of the tool providers have been transparent with you about
• 16:13
  SPEAKER_02: where they get their data and whether or not the providers consented
• 16:17
  SPEAKER_02: to it being taken and used for training?
• 16:20
  SPEAKER_00: I honestly haven't seen anything that's been very transparent.
• 16:24
  SPEAKER_00: I think the one time I've seen a few things is Meta sneakily changed their settings
• 16:30
  SPEAKER_00: so that your conversations or this or that are going to be used for AI.
• 16:34
  SPEAKER_00: I think TikTok did the same thing and we had to go turn it off, right?
• 16:37
  SPEAKER_00: But it was, I would say, a sneaky patch note, right?
• 16:40
  SPEAKER_00: They just kind of slipped it in and they didn't really say anything,
• 16:43
  SPEAKER_00: but they kind of said something.
• 16:45
  SPEAKER_00: But I haven't seen anybody really be super transparent where their data came from.
• 16:50
  SPEAKER_00: It kind of sucks.
• 16:51
  SPEAKER_00: And it also gives me a little bit of the ick as the kids say.
• 16:54
  SPEAKER_00: It's a little bit weird.
• 16:55
  SPEAKER_00: And I think even for TikTok, your videos can be used for AI, I believe.
• 17:00
  SPEAKER_00: And you individually have to go to every single video that you've created and turn it off.
• 17:06
  SPEAKER_00: Oh my goodness.
• 17:06
  SPEAKER_00: There isn't a mass update.
• 17:09
  SPEAKER_00: And so at this point, you're like, do I just delete it all?
• 17:13
  SPEAKER_00: They're also making it really hard for consumers to actually opt out.
• 17:17
  SPEAKER_00: It's kind of like canceling a subscription that they always hide it.
• 17:21
  SPEAKER_00: They make you go 30 different steps to go cancel something.
• 17:24
  SPEAKER_00: They should be probably cracking down on it,
• 17:26
  SPEAKER_00: like how they cracked down on canceling subscriptions.
• 17:29
  SPEAKER_00: We'll see how that goes.
• 17:30
  SPEAKER_02: Yeah, I'm not a TikTok user.
• 17:31
  SPEAKER_02: I never got comfortable with the way that they were going
• 17:34
  SPEAKER_02: to be handling people's information,
• 17:36
  SPEAKER_02: but I did not realize that you have to turn off for every individual video.
• 17:42
  SPEAKER_02: That's ridiculous.
• 17:44
  SPEAKER_00: Yeah, TikTok switched when the U.S. had to buy and they had to divest.
• 17:48
  SPEAKER_00: That was a whole other situation where I'm like, it's a little weird.
• 17:52
  SPEAKER_00: And then that whole AI model and that whole thing is I believe they're licensing out
• 17:57
  SPEAKER_00: the TikTok AI algorithm and then they just used it for us.
• 18:01
  SPEAKER_00: So it's a big kind of words.
• 18:03
  SPEAKER_02: Wow.
• 18:05
  SPEAKER_02: Yeah, having to opt out on everyone, that's not a customer.
• 18:08
  SPEAKER_02: Not great.
• 18:09
  SPEAKER_02: Yeah.
• 18:10
  SPEAKER_02: So you're in California,
• 18:11
  SPEAKER_02: which gives you somewhat more protection than the rest of us in other parts of
• 18:15
  SPEAKER_02: the states, which don't have that.
• 18:17
  SPEAKER_02: Not as much as people in Europe maybe, but more so than the rest of us.
• 18:20
  SPEAKER_02: Any of the new laws that have come out in California,
• 18:24
  SPEAKER_02: are they having an impact on your privacy in any ways that you can notice as
• 18:28
  SPEAKER_02: a resident?
• 18:28
  SPEAKER_00: Yeah, in a really good way, actually.
• 18:30
  SPEAKER_00: We just passed something about the data brokers.
• 18:33
  SPEAKER_00: Starting actually this month,
• 18:34
  SPEAKER_00: I already had put a request in the beginning of the year,
• 18:37
  SPEAKER_00: but there's websites like truepeoplesearch.com that have every single information
• 18:42
  SPEAKER_00: about you,
• 18:42
  SPEAKER_00: like your address, your email addresses, your phone numbers and everything.
• 18:47
  SPEAKER_00: And I had figured this out years ago, so I deleted all of my information.
• 18:51
  SPEAKER_00: But now there's one way,
• 18:53
  SPEAKER_00: one place in the state of California,
• 18:55
  SPEAKER_00: it's one of their official sites where you can just put in the request
• 18:58
  SPEAKER_00: and it sends it to all the data brokers and they have to delete all your data.
• 19:04
  SPEAKER_02: Oh, wow.
• 19:05
  SPEAKER_02: Yeah, which is really cool.
• 19:07
  SPEAKER_00: No, I think that's super cool for us because now no one can creepily stalk you on the internet.
• 19:12
  SPEAKER_00: And I think that was something that I thought about too with your data privacy.
• 19:16
  SPEAKER_00: When I found out my address was on the internet, I was like, this is scary.
• 19:20
  SPEAKER_00: Like if this goes to the wrong hands, someone can get swatted, right?
• 19:25
  SPEAKER_00: And so I'm really excited that we're kind of taking care of ourselves over here
• 19:28
  SPEAKER_00: and we're putting some data protection and privacy laws.
• 19:32
  SPEAKER_00: It's helping and that's one of the good things I'm seeing right now.
• 19:34
  SPEAKER_00: So yeah.
• 19:36
  SPEAKER_02: That's nice to hear about.
• 19:37
  SPEAKER_02: So you've built some AI tools yourself.
• 19:39
  SPEAKER_02: When you've built those tools,
• 19:41
  SPEAKER_02: I think I know what the answer might be,
• 19:44
  SPEAKER_02: but when you've worked with building these AI machine learning systems,
• 19:48
  SPEAKER_02: what can you share about the data that was used for it,
• 19:51
  SPEAKER_02: where it came from,
• 19:52
  SPEAKER_02: how it was obtained and whether or not consent was any issue with it?
• 19:56
  SPEAKER_00: Yeah, specifically a few jobs ago, we did it with our customer base, right?
• 20:02
  SPEAKER_00: And we had that data internally because it was like purchase history.
• 20:06
  SPEAKER_00: So that data came from us.
• 20:08
  SPEAKER_00: What never sat well with me was we did use a third party at one point
• 20:14
  SPEAKER_00: and it was called Newstar,
• 20:16
  SPEAKER_00: which was like This blew my mind as a 20 something year old at the time.
• 20:21
  SPEAKER_00: And it was the company who used to,
• 20:24
  SPEAKER_00: and I don't think the kids will ever know this,
• 20:25
  SPEAKER_00: but when you used to get like a cell phone service with one company and move it to
• 20:31
  SPEAKER_00: the other,
• 20:31
  SPEAKER_00: you would have to wait a few days to get transferred over to the new company.
• 20:37
  SPEAKER_00: And that company was Newstar.
• 20:39
  SPEAKER_00: I think they renamed themselves to Newstar.
• 20:40
  SPEAKER_00: And I don't think that they're named like that anymore.
• 20:43
  SPEAKER_00: But they had all your information.
• 20:45
  SPEAKER_00: And I remember working with them and they were like,
• 20:49
  SPEAKER_00: yeah, if you give us like, I forget what it was.
• 20:51
  SPEAKER_00: I think it might've been emails.
• 20:53
  SPEAKER_00: We can find out their location, this, that, blah, blah, blah.
• 20:58
  SPEAKER_00: And what segment of marketing that they fit personality profile wise.
• 21:02
  SPEAKER_00: And we'll send them flyers on your product.
• 21:06
  SPEAKER_00: That always tripped me out.
• 21:08
  SPEAKER_00: That did not sit well with me.
• 21:09
  SPEAKER_00: But that was also part of the data that we used
• 21:12
  SPEAKER_00: to build these customer retention analytics models, right?
• 21:15
  SPEAKER_00: And we had to figure out like why people were not buying our product or what they would like.
• 21:20
  SPEAKER_00: And there's even certain parts where We did some market research
• 21:23
  SPEAKER_00: with our own customers and we brought them in to do actual studies or we would surf
• 21:27
  SPEAKER_00: the internet on Reddit because people love talking about us on Reddit.
• 21:31
  SPEAKER_00: So it was just a mixture of things.
• 21:33
  SPEAKER_00: I think the most uncomfortable one was that and I would say at
• 21:36
  SPEAKER_00: the time it was Facebook ads and programmatic ads and Google Analytics
• 21:42
  SPEAKER_00: because we used to joke around and call it spyware.
• 21:45
  SPEAKER_00: Because it really was tracking every single movement but you are tracking
• 21:50
  SPEAKER_00: the movement of your consumer and figuring out why did they jump over here
• 21:54
  SPEAKER_00: and why did they come back or didn't come back right so there's
• 21:58
  SPEAKER_00: that other jobs were internal just customer data like when I used to work
• 22:02
  SPEAKER_00: in healthcare and that was pretty heavily locked down obviously there's a lot of PII
• 22:07
  SPEAKER_00: and HIPAA involved so you have to make sure that everything is anonymous right
• 22:11
  SPEAKER_00: but No concerns,
• 22:12
  SPEAKER_00: I would say,
• 22:12
  SPEAKER_00: even the industry that I am right now because, again,
• 22:15
  SPEAKER_00: everything's locked down and there's no PII involved and we're just looking
• 22:20
  SPEAKER_00: at metrics and efficiency.
• 22:22
  SPEAKER_02: Yeah, and when you're dealing with business to consumer types of activities
• 22:26
  SPEAKER_02: and obviously you're dealing with a lot of individual people,
• 22:29
  SPEAKER_02: business to business tends to be more locked down.
• 22:32
  SPEAKER_02: Yeah.
• 22:34
  SPEAKER_02: Yeah, that's good.
• 22:35
  SPEAKER_02: You mentioned that your data was out there in one of these data brokers
• 22:38
  SPEAKER_02: and you've cleaned it up.
• 22:39
  SPEAKER_02: Do you know any cases where your personal information has been used by someone
• 22:45
  SPEAKER_02: who bought it from a data broker, has been misused?
• 22:48
  SPEAKER_02: Any time like that where you know that your data has been used?
• 22:51
  SPEAKER_00: I don't think so, but I might have just been luck.
• 22:55
  SPEAKER_00: I will be honest.
• 22:56
  SPEAKER_00: I think it's just luck.
• 22:58
  SPEAKER_00: I've even had incidents where some of my stuff was stolen.
• 23:00
  SPEAKER_00: A lot of these data breaches and stuff, I thought.
• 23:02
  SPEAKER_00: Anthem had a data breach.
• 23:04
  SPEAKER_00: So many people had data breaches.
• 23:06
  SPEAKER_00: So I'm at the point where I don't think people have done anything negative to me
• 23:10
  SPEAKER_00: because I've also locked everything down.
• 23:12
  SPEAKER_00: You lock down your credit just so that no one can open something, right?
• 23:14
  SPEAKER_00: Just in case.
• 23:15
  SPEAKER_00: And there's so many data breaches so long.
• 23:18
  SPEAKER_00: Perpetually, you just got to lock everything down.
• 23:20
  SPEAKER_00: So I took more precautions than most people.
• 23:23
  SPEAKER_02: And at least in consumer situations,
• 23:25
  SPEAKER_02: we can usually choose whether or not to use that service, that tool.
• 23:29
  SPEAKER_02: As members of the public,
• 23:31
  SPEAKER_02: there are government systems and we honestly don't always have a choice.
• 23:34
  SPEAKER_02: Or if we travel, we run into TSA.
• 23:37
  SPEAKER_02: Have you had any experiences along those lines?
• 23:40
  SPEAKER_00: Yeah I didn't realize we could opt out with TSA with the photo recognition and
• 23:45
  SPEAKER_00: so I've been just doing it.
• 23:47
  SPEAKER_00: It still gets me a little apprehensive that my face is there.
• 23:51
  SPEAKER_00: I remember I came back from my trip from Europe and I have global entry and stuff
• 23:56
  SPEAKER_00: and I was walking into customs and they already knew my face
• 24:00
  SPEAKER_00: and it's like you can go right and it was like a two second thing and it was such
• 24:03
  SPEAKER_00: a weird thing for me where it was like You've only taken my photo once, right?
• 24:10
  SPEAKER_00: Or maybe a few times.
• 24:12
  SPEAKER_00: Where are you getting my face from?
• 24:13
  SPEAKER_00: Driver's license is probably from my 20s, right?
• 24:16
  SPEAKER_00: So where are you getting your data from?
• 24:17
  SPEAKER_00: I keep thinking, so you have a database somewhere.
• 24:20
  SPEAKER_00: Are you using different types of data brokers to just build a profile, right?
• 24:26
  SPEAKER_00: And so that's what I've been thinking about.
• 24:28
  SPEAKER_00: It's not great.
• 24:28
  SPEAKER_00: Now that I know I can opt out of the TSA thing, I'm going to stop doing that.
• 24:32
  SPEAKER_00: But yeah, it's been a little weird.
• 24:36
  SPEAKER_02: Yeah, I mean,
• 24:37
  SPEAKER_02: I haven't been through airports since they started doing that,
• 24:40
  SPEAKER_02: but a friend told me about the fact that signs in the airports,
• 24:44
  SPEAKER_02: the TSA signs in small print at the bottom that you could choose
• 24:48
  SPEAKER_02: to be manually verified if you prefer.
• 24:51
  SPEAKER_02: But some people don't feel comfortable doing that.
• 24:53
  SPEAKER_02: It's already not necessarily confrontational,
• 24:55
  SPEAKER_02: but it can be a stressful situation for a lot of people.
• 24:59
  SPEAKER_02: And some people just don't feel comfortable choosing that is unfortunate.
• 25:04
  SPEAKER_00: Yeah and I think it also makes you feel
• 25:06
  SPEAKER_00: a little bit weird like oh why wouldn't you right another case too is I don't think
• 25:11
  SPEAKER_00: people realize your iPhone your face ID I believe it's an internal ANN model
• 25:18
  SPEAKER_00: that you're just training over and over again whenever you're opening your phone.
• 25:23
  SPEAKER_00: That's another random thing that no one probably thinks
• 25:26
  SPEAKER_00: about either is your facial recognition,
• 25:27
  SPEAKER_00: your thumbprint,
• 25:29
  SPEAKER_00: Windows Hello right on your phone or your MacBook has a fingerprint sensor too.
• 25:33
  SPEAKER_00: It's like you're training some sort of AI.
• 25:37
  SPEAKER_00: Your biometrics are somewhere out in the universe.
• 25:41
  SPEAKER_02: Yeah, and the question is always whether or not that's really staying local or
• 25:44
  SPEAKER_02: if some information is being sent up to the mothership to be used for training and
• 25:49
  SPEAKER_02: for improving the models.
• 25:50
  SPEAKER_02: Like you said, it's probably buried in the terms and conditions.
• 25:54
  SPEAKER_02: They're not really written to be read, so I don't fault people for not reading them.
• 25:58
  SPEAKER_00: Yeah, I don't blame them either because I am guilty and I will not read that.
• 26:03
  SPEAKER_02: But yeah,
• 26:03
  SPEAKER_02: your point about the face ID,
• 26:04
  SPEAKER_02: I don't actually use it because I don't want my phone to be unlocked
• 26:08
  SPEAKER_02: without my true consent.
• 26:10
  SPEAKER_02: Like I said,
• 26:11
  SPEAKER_02: I haven't been traveling,
• 26:11
  SPEAKER_02: so it's not really a factor for me, but still I don't feel comfortable with it.
• 26:15
  SPEAKER_02: And it sounds like you don't either.
• 26:18
  SPEAKER_00: Yeah, no, I actually take it out whenever I travel.
• 26:22
  SPEAKER_00: So if I go through airports, I disable it.
• 26:24
  SPEAKER_00: I take it out.
• 26:25
  SPEAKER_00: I even put a longer passcode in and that's what I end up doing
• 26:28
  SPEAKER_00: because I didn't realize that we can just put the phone to your face and
• 26:31
  SPEAKER_00: then it's unlocked.
• 26:32
  SPEAKER_00: That's not great.
• 26:33
  SPEAKER_02: Yeah.
• 26:33
  SPEAKER_02: I think the rules are different around having permission to access your phone.
• 26:38
  SPEAKER_02: They can't make you give them a password, but they can put it up to your face.
• 26:45
  SPEAKER_02: People need to be aware of that.
• 26:48
  SPEAKER_02: Do you know any company that you intentionally gave your information to
• 26:52
  SPEAKER_02: that made you aware that they might use it for training AI or machine learning?
• 26:57
  SPEAKER_00: I think the biggest one was probably Meta on Instagram and they're using your chats
• 27:03
  SPEAKER_00: to train the model.
• 27:05
  SPEAKER_00: I opted out of it, but that was probably the first time that it was a little clear.
• 27:09
  SPEAKER_00: It was still hidden, but it was a little bit more clear.
• 27:12
  SPEAKER_00: I think WhatsApp was a little bit hidden too.
• 27:15
  SPEAKER_00: I noticed that if you don't go into the chat settings of that specific chat
• 27:20
  SPEAKER_00: and you don't put advanced chat protection or something like that or toggle it on,
• 27:24
  SPEAKER_00: it will use your conversations to train the model.
• 27:27
  SPEAKER_00: So I'll go through the little settings here and there,
• 27:30
  SPEAKER_00: but those are the two,
• 27:32
  SPEAKER_00: and then both Meta,
• 27:33
  SPEAKER_00: those are the two things that I've actually seen that they're
• 27:36
  SPEAKER_00: a little bit transparent about it.
• 27:38
  SPEAKER_00: I haven't seen anything else being super transparent.
• 27:41
  SPEAKER_02: You mentioned Meta and Instagram.
• 27:42
  SPEAKER_02: I never used Instagram because I just wasn't comfortable with them.
• 27:46
  SPEAKER_02: And two years ago when Meta came out and said,
• 27:48
  SPEAKER_02: we're going to start using your pictures and your chats and everything.
• 27:51
  SPEAKER_02: It's like, no.
• 27:53
  SPEAKER_02: I tried to opt out.
• 27:54
  SPEAKER_02: People in Europe, their opt outs got respected.
• 27:56
  SPEAKER_02: Us was like, yeah, we don't have to, so we're not going to.
• 27:59
  SPEAKER_02: Okay.
• 28:00
  SPEAKER_02: So I went ahead and deleted pretty much everything I'd ever put in there.
• 28:04
  SPEAKER_02: Not that they didn't keep it and use it anyway.
• 28:06
  SPEAKER_02: There's no way to know, but at least I took it out.
• 28:08
  SPEAKER_02: I remember hearing from other people who were on Instagram
• 28:11
  SPEAKER_02: that they got concerned more recently when there was another, yeah,
• 28:14
  SPEAKER_02: we're going to start using your pictures and using your information from Instagram.
• 28:18
  SPEAKER_02: I know some people use it for their businesses and it's hard to leave.
• 28:22
  SPEAKER_00: Yeah I know it's super hard to leave and even if you think about it right they have
• 28:28
  SPEAKER_00: the biggest market share so anything that you do any type of advertising or even
• 28:34
  SPEAKER_00: if you are like a politician for example and you're trying
• 28:38
  SPEAKER_00: to reach your constituents right people are on threads people are on TikToks people
• 28:43
  SPEAKER_00: on Facebook there are different demographics that
• 28:46
  SPEAKER_01: people use different types
• 28:47
  SPEAKER_01: of platforms so if you're not hitting
• 28:49
  SPEAKER_00: all of them you're going to lose a lot
• 28:51
  SPEAKER_00: of traction so it kind of sucks and it puts I would say the little people in
• 28:56
  SPEAKER_00: a pretty big predicament of damned if you do and you damned if you don't and
• 28:59
  SPEAKER_00: so how do you approach it in a way where you're limiting the stuff that you put on
• 29:04
  SPEAKER_00: that it can't be really used for their benefit right some things I've been thinking
• 29:08
  SPEAKER_00: about lately yeah
• 29:11
  SPEAKER_02: those are all good points like you said some people have
• 29:16
  SPEAKER_02: to be there but Trying to be careful about what they post is obviously a concern.
• 29:20
  SPEAKER_02: So last question and then we can talk about anything else that you want.
• 29:24
  SPEAKER_02: The public distrust of AI and tech companies has been growing.
• 29:29
  SPEAKER_02: There's been even a backlash recently.
• 29:31
  SPEAKER_02: You mentioned meta with what some people are calling now the pervert glasses.
• 29:35
  SPEAKER_02: Yeah.
• 29:37
  SPEAKER_02: Yeah, that's basically spyware.
• 29:39
  SPEAKER_02: So there's been some backlash.
• 29:41
  SPEAKER_02: There's been a growing sense that the companies just can't be trusted.
• 29:44
  SPEAKER_02: Move fast and break things is one thing,
• 29:46
  SPEAKER_02: but actually putting spyware tools like that into the hands of people
• 29:50
  SPEAKER_02: who might not do good things with them is really concerning.
• 29:55
  SPEAKER_02: I'm wondering if you think that any of these big AI companies,
• 30:00
  SPEAKER_02: big tech companies can be trusted and what it would take for you to trust them,
• 30:05
  SPEAKER_02: or if that's really not even possible?
• 30:09
  SPEAKER_00: It's a really tough question, but I don't think they are to be trusted.
• 30:13
  SPEAKER_00: I think the one thing they could do is being transparent where their data comes
• 30:17
  SPEAKER_00: from, but they're not going to.
• 30:20
  SPEAKER_00: I mean,
• 30:21
  SPEAKER_00: we have the incident with Claude right now who is actively buying historical books
• 30:27
  SPEAKER_00: and one-off editions.
• 30:28
  SPEAKER_00: They're rare books and there's only like one or two of them in
• 30:31
  SPEAKER_00: the whole entire world and they're feeding them to their model and then destroying
• 30:35
  SPEAKER_00: the books.
• 30:36
  SPEAKER_00: They're destroying history and art.
• 30:38
  SPEAKER_00: I don't know how you come back from that.
• 30:41
  SPEAKER_00: And then especially these big corporations with the data centers, right?
• 30:46
  SPEAKER_00: They're actively destroying communities at this point.
• 30:49
  SPEAKER_00: They come back from it when they start realizing the issues that they're causing
• 30:55
  SPEAKER_00: in terms of destroying history or destroying environments.
• 30:58
  SPEAKER_00: And then actually showing the transparency of data and where everything is coming from.
• 31:03
  SPEAKER_00: Because I think once you lay it out on the table and show it to your consumers,
• 31:08
  SPEAKER_00: They're not going to have complete trust in them, but at least there's transparency.
• 31:12
  SPEAKER_00: And I think that's a step towards the right direction.
• 31:15
  SPEAKER_00: But again, I think a lot of these companies are just margin and bottom line.
• 31:21
  SPEAKER_00: So I don't know if they're going to be actively doing that.
• 31:23
  SPEAKER_00: I don't know if they actively care, right?
• 31:25
  SPEAKER_00: Because they have these large contracts with big companies like Microsoft.
• 31:29
  SPEAKER_00: Microsoft has Claude backing co-pilot.
• 31:32
  SPEAKER_00: They're not going to really hurt if a consumer isn't using their AI tool.
• 31:37
  SPEAKER_00: Especially they embedded in something that's so big like Co-Pilot that's on Microsoft.
• 31:41
  SPEAKER_00: I do think that if they show the data lineage and transparency in that and
• 31:45
  SPEAKER_00: where they got it from,
• 31:46
  SPEAKER_00: people would probably understand, oh, this came from here, right?
• 31:49
  SPEAKER_00: This is credible.
• 31:52
  SPEAKER_02: Yeah.
• 31:52
  SPEAKER_02: So out where you are in Los Angeles,
• 31:54
  SPEAKER_02: obviously concerns about the environment,
• 31:57
  SPEAKER_02: concerns about water supplies and such are huge.
• 31:59
  SPEAKER_02: And then there was the impact on power systems.
• 32:02
  SPEAKER_02: Are you seeing or hearing anything in your area?
• 32:06
  SPEAKER_02: Where people were resisting or pushing back on data centers being located in those areas?
• 32:11
  SPEAKER_00: Oh yeah, we actually just won a couple.
• 32:14
  SPEAKER_00: The city of Monterey Park,
• 32:16
  SPEAKER_00: I believe it's Monterey Park or Montevideo,
• 32:18
  SPEAKER_00: it's one of the two,
• 32:19
  SPEAKER_00: but I think it's Monterey Park, banned data centers from the city.
• 32:23
  SPEAKER_00: So we beat them on that.
• 32:25
  SPEAKER_00: We were going to have a data center in El Segundo.
• 32:27
  SPEAKER_00: That's near LAX airport.
• 32:30
  SPEAKER_00: And the guy just walked it back.
• 32:32
  SPEAKER_00: They got so much public backlash.
• 32:34
  SPEAKER_00: They're like, we're not building it.
• 32:35
  SPEAKER_00: The actual developers said, nope, we're just going to respectfully not do this.
• 32:39
  SPEAKER_00: And so we've been pushing back a lot just because if you look at
• 32:43
  SPEAKER_00: where they're putting the data centers at across the country,
• 32:47
  SPEAKER_00: It's always in lower socioeconomic areas
• 32:51
  SPEAKER_00: and they're always polluting all those areas because at the end of the day,
• 32:55
  SPEAKER_00: it's late and it's terrible to sound.
• 32:57
  SPEAKER_00: They're not going to put it where there's off-wind people are.
• 33:01
  SPEAKER_00: They're going to put it where they think people aren't going to be able to push back.
• 33:05
  SPEAKER_00: So we've seen people push back and they're protesting and they're getting
• 33:10
  SPEAKER_00: with their local Congress people and their senators and holding them accountable.
• 33:15
  SPEAKER_00: They're going with their city council members and showing up to the meetings
• 33:19
  SPEAKER_00: and making sure that if it hasn't been done yet,
• 33:23
  SPEAKER_00: they're trying to corner them and say, nope.
• 33:26
  SPEAKER_00: We're not going to allow this.
• 33:27
  SPEAKER_00: So I'm happy about that because there has to be a better way
• 33:30
  SPEAKER_00: and we shouldn't be destroying the environment,
• 33:33
  SPEAKER_00: destroying the communities in the name of business.
• 33:36
  SPEAKER_00: That's something ethically I can never do and there are places I grew up in.
• 33:40
  SPEAKER_00: Around so it feels bad.
• 33:42
  SPEAKER_00: When I drive to places where I grew up and I see a giant data center,
• 33:46
  SPEAKER_00: that's why I tell myself I don't want to be part of the problem either.
• 33:49
  SPEAKER_00: That's why I personally will limit my use of these big systems and all that kind
• 33:53
  SPEAKER_00: of stuff because I'm going to be a hypocrite if I continue to do this.
• 33:57
  SPEAKER_02: Yeah I mean on the one hand it's hard to avoid AI but where we have a choice and
• 34:01
  SPEAKER_02: where we have options where a human could draw the art instead of pushing
• 34:05
  SPEAKER_02: a button 20 times in a chat and you still don't care what you want
• 34:08
  SPEAKER_02: and you've just wasted all that energy and water.
• 34:13
  SPEAKER_02: We do have some choices.
• 34:14
  SPEAKER_02: That's my perspective on what I call everyday ethical AI.
• 34:17
  SPEAKER_02: It's easy to feel powerless in front of the big tech companies, but we do have some power.
• 34:22
  SPEAKER_02: We have some agency,
• 34:23
  SPEAKER_02: not just in pushing back with our communities on data
• 34:27
  SPEAKER_01: centers and
• 34:27
  SPEAKER_01: on what companies are doing with our data, but just in our day-to-day choices.
• 34:31
  SPEAKER_02: And I always encourage people to keep that perspective and think
• 34:36
  SPEAKER_02: about what they can do,
• 34:37
  SPEAKER_02: what they want to use AI tools for and why,
• 34:41
  SPEAKER_02: and what they don't want to use it for and why.
• 34:44
  SPEAKER_02: I call it an AI policy.
• 34:45
  SPEAKER_02: That sounds very formal,
• 34:47
  SPEAKER_02: but it's really just how do you make your decisions day to day about what you do
• 34:51
  SPEAKER_02: and don't use, and how does that line up with your values?
• 34:54
  SPEAKER_02: And that's something we can all do.
• 34:57
  SPEAKER_00: No, absolutely,
• 34:58
  SPEAKER_00: because I think that's a really good way to say it, like the AI policy for yourself.
• 35:02
  SPEAKER_00: I guess that's what I've been trying to do unofficially is what am I willing
• 35:06
  SPEAKER_00: to give up like the ease of access through
• 35:08
  SPEAKER_00: an AI tool versus what are some things I can't give up or I have to leverage
• 35:13
  SPEAKER_00: or can I turn off the AI overview on Google search, right?
• 35:18
  SPEAKER_00: Just these small little quality life kind of things,
• 35:21
  SPEAKER_00: right, that we can get rid of or just include.
• 35:24
  SPEAKER_02: Yeah.
• 35:24
  SPEAKER_02: And you can turn it off.
• 35:25
  SPEAKER_02: I've told some people about the UDM 14 shortcut and most people haven't ever heard of it.
• 35:30
  SPEAKER_02: It's like, I can turn that off?
• 35:31
  SPEAKER_02: Oh, well, keep it turned off or I use a different browser where that's not the default.
• 35:37
  SPEAKER_02: It's just a waste.
• 35:38
  SPEAKER_02: I mean, if I'm searching for something, most I know what I'm looking for.
• 35:40
  SPEAKER_02: I want the company's website and I want their support page.
• 35:44
  SPEAKER_02: I want the technical information, the specs on this device.
• 35:46
  SPEAKER_02: And I don't need a generated answer for that.
• 35:49
  SPEAKER_00: And sometimes that generated answer is wrong too.
• 35:52
  SPEAKER_00: Yes.
• 35:53
  SPEAKER_00: So I'll just click on the website and figure out the tech specs of some things.
• 35:56
  SPEAKER_02: Well, it's great that you're being so thoughtful about how you use it.
• 35:59
  SPEAKER_02: It's hard to avoid and there can be benefits to it,
• 36:02
  SPEAKER_02: but being careful and using what some people call our agency.
• 36:07
  SPEAKER_02: We do have some power in that.
• 36:08
  SPEAKER_02: It's really been fun talking with you.
• 36:10
  SPEAKER_02: I'm glad you're doing it and I appreciate you making the time for the interview with me.
• 36:13
  SPEAKER_02: Any other thoughts you have about AI?
• 36:15
  SPEAKER_02: Anything you want to share with our audience?
• 36:17
  SPEAKER_00: No, I think we hit it all.
• 36:19
  SPEAKER_00: Especially we talked a little bit about data centers and stuff.
• 36:21
  SPEAKER_00: So thank you so much.
• 36:23
  SPEAKER_00: This was very fun.
• 36:25
  SPEAKER_01: About this interview series and newsletter.
• 36:29
  SPEAKER_01: This post is part of our AI6P interview series on AI software and wetware.
• 36:36
  SPEAKER_01: It showcases how real people around the world are using their wetware,
• 36:40
  SPEAKER_01: brains and human intelligence,
• 36:42
  SPEAKER_01: with AI-based software tools who are being affected by AI.
• 36:47
  SPEAKER_01: And we're all being affected by AI nowadays in our daily lives, perhaps more than we realize.
• 36:53
  SPEAKER_01: For some examples, see post, but I don't use AI.
• 36:58
  SPEAKER_01: We want to hear from a diverse pool of people worldwide in a variety of roles.
• 37:03
  SPEAKER_01: If you're interested in being a featured interview guest,
• 37:06
  SPEAKER_01: anonymous or with credit, please get in touch.
• 37:12
  SPEAKER_01: 6 P's in AI Pods AI6P is a 100% reader-supported publication.
• 37:19
  SPEAKER_01: All new posts are free to read and listen to.
• 37:22
  SPEAKER_01: To automatically receive new AI6P posts and support our work, consider becoming a subscriber.
• 37:29
  SPEAKER_01: It's free.