132. Lyudmila Lugovskaya: The Hidden Costs of Creative Automation
In this episode of Women WithAI, Dr. Lyudmila Lugovskaya brings her extensive experience in AI and data science to our conversation. We discuss how she helps companies navigate the complex landscape of generative AI to achieve real business outcomes. Lyudmila highlights the common misconceptions surrounding AI, particularly the overestimation of its capabilities and the importance of having clean, organised data. We explore the evolving job market as AI becomes more integrated into various industries, and the necessity for humans to adapt and maintain essential skills. Our discussion emphasises the potential of AI as a tool for innovation while also recognising the challenges and responsibilities that come with its use.
Takeaways:
- Dr. Lugovskaya emphasises the importance of having clean and organised data for successful AI implementations.
- Companies often have unrealistic expectations about AI, believing it can instantly solve their business problems.
- As generative AI rapidly evolves, continuous learning and adaptation are essential for professionals in the field.
- The conversation highlights that while AI can automate tasks, it also requires human oversight and input for effective results.
- Lyudmila suggests starting to use AI tools with detailed prompts to achieve better outcomes and efficiency.
- The emergence of AI is likely to change job roles, creating new opportunities while rendering some tasks obsolete.
Links referenced in this episode:
Transcript
Foreign you're listening to with aifm.
Speaker A:Hello and welcome to Women with AI, the podcast dedicated to amplifying the voices and perspectives of women in a world filled with artificial intelligence.
Speaker A:My guest today has a vast amount of experience across different industries and helps companies navigate the complex world of data and generative AI to deliver tangible business outcomes.
Speaker A:But before we start the podcast, let me tell you a little bit about her.
Speaker A:Dr.
Speaker A:Lyudmila Lugovskaya is an AI and data science consultant.
Speaker A:After a PhD at Cambridge, she moved from credit analysis at bank of America to data analytics at Lloyds Banking Group, where she worked on early data science and responsible AI projects.
Speaker A:Lyudmila went on to build machine learning models, risk algorithms, and generative AI tools across tech, maritime, intelligence, and sustainability industries.
Speaker A:Today, she works independently, helping companies Design and deliver AI systems that solve real business problems.
Speaker A:Dr.
Speaker A:Lyudmila Lugovskaya, welcome to Women with AI.
Speaker B:Hello.
Speaker B:Thank you very much for having me.
Speaker A:It's great to have you here.
Speaker A:So interesting because you started in business and management before discovering your passion for data science.
Speaker A:I guess.
Speaker A:So.
Speaker A:Can you share with us how that journey unfolded?
Speaker B:Yes, certainly.
Speaker B:So I did a PhD at Cambridge in small business finance, and as part of my thesis, I was working on a quantitative study trying to predict the probability of bankruptcy based on some financial ratios.
Speaker B:And we would say today that it's a classical binary classifier problem in machine learning.
Speaker B:Back then we were just calling it quantitative methods.
Speaker B:But when data science started to emerge as its own field, I immediately recognized that this is similar to what I did in my PhD.
Speaker B:That, that was the, the bit I really enjoyed and I kind of wanted to, to do more of that.
Speaker B:And it was also attractive that data science is an applied discipline.
Speaker B:So, you know, I was getting the best of both world worlds.
Speaker B:You know, the, the curiosity of studying something which I enjoyed in academia, but also seeing how it can be applied and make a difference.
Speaker A:Okay.
Speaker A:Because it's like, it feels like AI is an emerging technology, but I guess as you say, it's, it's, you know, what you were doing is predicting things.
Speaker A:So really it's, you know.
Speaker A:Oh, it's been around for ages, hasn't it, really?
Speaker A:I guess you spotted.
Speaker B:Yeah, I suppose we can say that generative AI LLMs are an emerging technology, but AI as a broader discipline has been around for some decades.
Speaker B:And 10 years ago, when I was kind of starting in the field, it was all about data science and machine learning, which is also a branch of AI.
Speaker B:So that, that's kind of how my journey started there.
Speaker A:And so you spent, I guess, almost the last decade working with AI in finance, and then more recently, as you say, you're working independently, supporting clients, as a data science freelancer.
Speaker A:How does that sort of like, I don't know, what sort of challenges and opportunities do you see as sort of, as you say, generative AI is sort of gaining traction?
Speaker B:Yeah.
Speaker B:I mean, during all my time in data, I think one thing that is constant is the change.
Speaker B:Because when I first joined the discipline, everything was very new.
Speaker B:There was a lot to learn, a lot to absorb, and it always felt like, oh, I'm still not quite there.
Speaker B:There is this algorithm I need to learn, there is this new framework, there is this package.
Speaker B:And then I just accepted that this is very much part of the profession, part of the field.
Speaker B:It's very dynamic.
Speaker B:It's almost impossible to feel like you are caught up on everything because, you know, new, new things are emerging all the time.
Speaker B:And with Genai, I think we see even more of that.
Speaker B:If, you know, when I was studying in data science, maybe every few months there would be something kind of big happening, or every year or so, now it's almost every week, if not more frequently we see some new LLM appearing and everyone is excited.
Speaker B:There is a new framework.
Speaker B:And yeah, I think that's just the way it is.
Speaker B:I think it's impossible to be on top of everything, but kind of knowing what's important and just trying to follow the field and then learn things as they become needed in practical projects is, I suppose, how to, to manage that and how to stay sane.
Speaker B:So at least that's been my approach.
Speaker A:How do you stay on top of everything that's happening?
Speaker A:Do you sort of sign up to all the millions of newsletters that are out there, or do you have a kind of like a secret in with what's going on in the AI industry?
Speaker B:As I say, I, in all honesty, I don't think I'm staying on top of everything because it's not possible.
Speaker B:So, yes, I sign up to some newsletters and I listen to podcasts and there are educational videos I try to watch, and there are various people I follow.
Speaker B:But at the end of the day, I still know that there are many interesting things that I miss that I don't see, that I don't have time for.
Speaker B:And then it's, as I've said on the need to basis, when I'm working on a project and I feel I need to learn about something specific, then I, I go and learn about it in a kind of more targeted manner.
Speaker A:When you're working with clients, because I guess some of them are probably just beginning their AI journey.
Speaker A:You know, do you.
Speaker A:Do they are sort of like common misconceptions that people have, or like the big ones that sort of people have.
Speaker A:Like, how do you guide people through what's happening?
Speaker B:Yeah, I think that there are very high expectations of AI, and that's something I previously saw with machine learning.
Speaker B:So sometimes the companies think that it's almost a magic wand which will somehow transform their business overnight.
Speaker B:All you need to do is to, I don't know, hire someone who is doing AI and build a few things and then it will immediately bring those huge benefits.
Speaker B:I think it doesn't always happen so easily.
Speaker B:So whilst indeed the potential of technology is enormous, there is so much we can do with it.
Speaker B:I think the foundations are still very much the same as they were with machine learning, as they were with data science.
Speaker B:It's basically having clean, good, organized data.
Speaker B:And that's something that companies don't often have.
Speaker B:Either the data is just not accessible, there isn't enough of it, or there are some siloed data sources, they're not a very good quality, they're not easy to work with, and so on.
Speaker B:And then the first thing that is very important to do is to basically build out that foundation.
Speaker B:So I think it was true 10 years ago, five years ago, and it's still very, very true now.
Speaker B:And then the second kind of common pitfall I see is that there is again this excitement.
Speaker B:It's like, okay, AI will solve all my problems, but what specific problems?
Speaker B:So there is often not enough concrete definition of what exactly we're trying to achieve, and then the standing of what we can and can't do with AI.
Speaker B:So one of the things I usually recommend to companies is to slow down and really try to figure out what it is that they want to do and to set more realistic expectations.
Speaker B:And that often means things will take longer than people think and the results may be not, like, as amazing as they expect, but still very worthwhile.
Speaker B:So I think, you know, once that understanding is established, usually the projects go much, much better.
Speaker A:So I guess, yeah, you're right.
Speaker A:You're the magician with the magic wand of AI to come and help them.
Speaker A:And you're right, I hadn't really thought about that because I think you're right.
Speaker A:Lots of people just think, oh, well, it'll help solve all these problems, or it will help make life easier.
Speaker A:But you still need to know what is the problem you're trying to solve?
Speaker A:Where are you trying to get to?
Speaker A:How are you going to do that?
Speaker A:So it's still.
Speaker A:There's a lot of work that has to be done by us, isn't there?
Speaker A:You know, by humans in the loop.
Speaker B:Yeah, exactly.
Speaker B:And then also I think it's very important to pick the right tool for the, for the job.
Speaker B:So sometimes, you know, a certain model or a certain technology becomes fashionable and then everyone think that it needs to be applied to every problem.
Speaker B:And that's not always the case.
Speaker B:It may be true.
Speaker B:But before just making an assumption, we, we need, you know, the latest machine learning algorithm or the latest gen AI.
Speaker B:It's still worth slowing down and asking what exactly are we trying to do?
Speaker B:Is there a simpler method than can probably achieve the same?
Speaker B:Sometimes there is some, sometimes there isn't.
Speaker B:But I wouldn't always rule out the simpler methods, simpler models first because although intellectually it is more exciting to play with the latest tech, from the practical point of view, it may or may not be the best solution.
Speaker B:And that's, I think that is just something to be, to be aware of.
Speaker A:It's not always the shiny new toy that's going to.
Speaker B:Yeah, yeah, exactly.
Speaker B:I think that results.
Speaker B:That's very much that effect.
Speaker B:Yeah, that aspect.
Speaker A:Because going back to what you said about when you were before, when you were predicting bankruptcy, like how, I mean, are people using AI to do the same thing that you were doing?
Speaker A:Or was, as you say, was it already AI that you were using?
Speaker B:Well, it depends on how you want to define AI.
Speaker B:Right.
Speaker B:So I think currently when we speak about AI, people often assume generative AI, so it definitely wasn't that.
Speaker B:But if we take the broader definition of AI, which existed at least since the middle of the 20th century, then yes, I actually prefer to use a broader definition of AI.
Speaker B:I think the AI hasn't appeared two years ago when generated AI models have started becoming popular.
Speaker B:It was building up on other work and other foundations.
Speaker B:So when I was predicting bankruptcy, I was using some binary classifiers and that's part of machine learning and machine learning is part of AI.
Speaker B:So from that point of view, I think it's fair to say that yes, I was using AI back then, but at the end it's the matter of which terminology we want to use and.
Speaker A:Talk about sort of terminology and that kind of stuff.
Speaker A:Because when we spoke previously, you raised such a powerful question that I've been thinking about ever since, which was what happens to the job market and the distribution of labor with AI because it's changing the way that people do things, the jobs that people can do.
Speaker A:So what are your thoughts on kind of how, what, as I say, what we would give to AI, or what should we keep for ourselves or what kind of new jobs should we be thinking about?
Speaker A:What do you think?
Speaker B:Yeah, I think that there is a lot of anxiety currently because AI is probably a new industrial revolution.
Speaker B:And each industrial revolution, I think, has been bringing its own set of challenges, anxieties and opportunities.
Speaker B:So for example, in the first industrial revolution, which kind of happened, I think in their early 80s, 19th century, we lost some artisan jobs, but then many manufacturing jobs, mechanical engineers started appearing.
Speaker B:And then the second industrial revolution again, some, some manual jobs were lost.
Speaker B:But then electrical jobs, electrical engineering appeared and more research and development jobs also became in demand.
Speaker B:And probably something similar will be, will be happening now.
Speaker B:I think there is no doubt that AI will be changing the job market, but it doesn't necessarily mean that it will be just replacing humans and humans will not be needed.
Speaker B:I think it just means that our roles will be different.
Speaker B:Some of the jobs that exist will disappear, some new ones will appear and we just need to adapt to that.
Speaker B:So I think that's the good news.
Speaker B:The bad news is that again, I think there is often this excitement that once we automate things, we will just have so much free time.
Speaker B:But that's not what's been happening historically, I think.
Speaker B:So, for example, when the washing machines first appeared, some people were saying, well, now women, because at the time it was mostly women doing household tasks.
Speaker B:They will have so much more free time to just enjoy life.
Speaker B:But that's not quite what happened.
Speaker B:I think the standards changed, so now that washing became easier, people were expecting cleaner clothes or change of clothes more frequently.
Speaker B:And that didn't result in free time, it just resulted in new expectations.
Speaker B:So it's.
Speaker B:It seems likely that AI will be just the same that things which are currently taking a long time, we will be able to do them much faster, but that will become a new norm.
Speaker B:So whether we will get that, that time back, I, I'm not sure, but it seems to me that probably not in the near term, maybe in some more remote future.
Speaker A:Because you want AI, as you say, to be sort of taking or giving us more time, but say it's, well, we're all going to have to spend time learning how to use it anyway and doing that.
Speaker A:But yeah, I hadn't thought about that yet.
Speaker A:If it's easier to wash things, then yes, you're going to wash your clothes more often.
Speaker A:I haven't thought about that.
Speaker B:Yeah, of course in the more kind of remote future, but again, I don't think it, it, it's likely to happen in the next like 10 years or so, maybe, maybe later when AI is so powerful and it genuinely can do so many of the jobs that people are doing that even with new jobs we just don't need as many, as many humans.
Speaker B:Then the question becomes how does economy work in general?
Speaker B:How does the distribution of wealth work?
Speaker B:I think it's a big societal and philosophical question, so that's kind of beyond my narrow area of expertise.
Speaker B:But I think one of the interesting debates that's happening is whether the universal basic income will become not just a nice idea, but a necessity.
Speaker B:And also to what extent extent should the benefits of AI be private as opposed to public?
Speaker B:Especially given that LLMs are being trained on very much, you know, the, the shared body of knowledge.
Speaker B:Then there is an argument that is only fair to distribute the, the benefits also more evenly across the society.
Speaker B:But I think we're still far from figuring this out.
Speaker B:This is like one of the big questions the society will, will have to face and find an answer to in the coming decades.
Speaker A:Because I saw something earlier and apologies if everyone else knows this, I haven't looked at it, but apparently the new Pope has already made a statement about how we know we have to be careful about how we're using AI and kind of like all the ethics that come into it and, and that kind of thing.
Speaker A:So I don't know.
Speaker A:I mean, how do you, do you think it.
Speaker A:There's, there's an easy way to sort of make sure that AI remains safe and reliable and ethical.
Speaker B:Well, I certainly don't think that it's easy.
Speaker B:It's a very important task that we'll need to tackle.
Speaker B:But there are many aspects of it from practical of how do we make sure the AI doesn't do anything harmful, malicious, how do we make sure that their goals are aligned with humanity and we don't get those very smart robots with their own aims.
Speaker B:But that's a bit of the science fiction, but probably more realistically and closer to home is what happens with jobs, what happens with profits?
Speaker B:So if we have huge automation and does many businesses become more profitable, then how, how do we share it across, across society?
Speaker B:So I think these are some of the important questions.
Speaker B:But the good news is then again, I think the small guys, it's not just, you know, the huge corporations which, which, which can benefit because Many things are becoming so much easier.
Speaker B:So I think we will also see the rise of solopreneurs where you are just a one man band or a very small company and where previously you would need many people and thus a significant capital, now with the help of AI, you can just build your own website, your own e commerce shop, your own accounting this or that.
Speaker B:So I think we're just at the start of it.
Speaker B:But the potential for automation and for doing more with less resources is definitely there.
Speaker B:So I think that's probably a more kind of optimistic angle to look at all of that.
Speaker B:So there is a lot of opportunities.
Speaker A:For everyone because it feels like that we should be.
Speaker A:You've mentioned before the challenge of cognitively demanding work and kind of, where do you kind of ring fence that?
Speaker A:Because I'm sort of thinking creatively, like writing and design.
Speaker A:We probably don't want to outsource all of that to AI or some people won't because those skills, if we lose those skills, is everything going to look the same?
Speaker A:Do you think we can stay creatively sharp?
Speaker B:Yeah, to be honest, this is one of the questions that has been worrying me somewhat as well because I just notice in my own work even that sometimes I'm becoming a bit lazy.
Speaker B:Things I normally would, would do on my own.
Speaker B:I then ask an LLM and then I'm thinking, why have I just done it?
Speaker B:I could have done it myself.
Speaker B:But I think there are a couple of things which still make me think that we will retain some of those complex and creative tasks for ourselves, one way or the other.
Speaker A:So.
Speaker B:So I think first the content generated by AI will start looking increasingly the same because all LLMs are trained on the same, more or less within some differences on the same purpose of text.
Speaker B:And some of the content is really similar.
Speaker B:So if you've seen enough of it, you can already just have a quick glance and you know it's been written by AI and it's kind of a bit boring.
Speaker B:So there will be a point of situation where no one will read that and the originality will be yet again in demand and people, people will, will start doing it because otherwise no one, no one will read, no one will care.
Speaker B:Also, I think that to be effective guides for AI to be able to guide their work, assess their work, we still need to have the skills ourselves because maybe not all low level skills, but at least some of the key skills that we are asking the AI to do.
Speaker B:It's important for people to retain.
Speaker B:And also I think that there will be, currently we have like prompt Engineer, AI engineer.
Speaker B:But I think as the field is developing, we will have more specialized professions that one way or another concern interacting with AI.
Speaker B:So maybe AI human design or ethics, or building architecture of many agents and coordinating them.
Speaker B:So one way or another it will require good prompting.
Speaker B:And good prompting means good writing and good writing means clear thinking.
Speaker B:So I think that if, if we outsource many things, we still need to, to be regularly writing ourselves because if anything it will be only gaining prominence and even a bigger role in the.
Speaker A:AI world completely right there.
Speaker A:Because I think lots of people are saying, oh, AI is not going for jobs.
Speaker A:It's, you know, people that know how to use AI.
Speaker A:But I, I do see that it might go in a cycle and we might just all get really lazy and let it do it.
Speaker A:And then I think we'll just say, hang on, everything looks the same.
Speaker A:No, we need that creativity back.
Speaker A:And hopefully there's people that have been honing those skills still and doing it because I, I do it, I ask, you know, chat, GPT or copilot or whatever to do something really quickly.
Speaker A:And then I do start to question what would I have been able to do that myself?
Speaker A:What if I've lost the skill of how to kind of like reword that or rephrase it?
Speaker A:So, yeah, I think that's really interesting.
Speaker A:I think that's the kind of positive, you know, good news story, hopefully that.
Speaker B:I, I hope so.
Speaker B:Yeah.
Speaker B:Maybe it won't be in, you know, all things we currently do ourselves and not even all types of writing, for example.
Speaker B:So there is a lot of like transactional writing where, you know, we're just trying to convey some message.
Speaker B:It's not particularly interesting or creative, we just need to do that.
Speaker B:But maybe for those types of interactions it will be AI to AI.
Speaker B:Maybe there won't be even many humans reading that type of text, but the types of texts where there are some new ideas or fiction or whatever else, some sort of really interesting literary text or technical text with truly original ideas, I think we will still appreciate when they're written by humans and we will need to maintain that skill to stay relevant.
Speaker B:That's my hope.
Speaker A:Yeah, that's my hope too, because otherwise we're on the edge of this precipice and everything is just going to become the same because it's been done by AI and then we won't know.
Speaker A:Yeah, because you said like, go.
Speaker A:You know, it's not just big companies, it's really useful for sort of smaller entrepreneurs or solopreneurs.
Speaker A:And do you see sort of other freelancers and sort of consultants?
Speaker A:Are they sort of doing what you're doing?
Speaker A:Is what is the work evolving for them?
Speaker B:Yeah, I think that everyone is using AI one way or another.
Speaker B:So in my work I often use it, for example, for brainstorming or showing some parts of the code that I'm writing and asking for opinion for how to optimize it.
Speaker B:And often when I look at how much I have written, how much back and forth there have been, I'm thinking, well, I haven't necessarily saved time, but I clarified some of my thoughts.
Speaker B:So that's kind of where I see the value.
Speaker B:So something where I probably would spend time just reading, searching for information or trying out some ideas, failing before it works.
Speaker B:I feel that that kind of helps me cut through that, do some back and forth with AI and get some, some results or some preliminary results.
Speaker B:And then when I'm ready to discuss it with the team, I already show them something a bit more mature than I would have done without using AI.
Speaker B:So I think that other freelancers, from what I can tell, are kind of doing the same.
Speaker B:They get a lot of the groundwork done with AI.
Speaker B:They're doing some drafting, whether it's text or code or proposal.
Speaker B:Overall, I think it makes things more efficient because the human input is still very valuable, still very required.
Speaker B:But you just do it at a later stage.
Speaker B:Before it could be draft one, draft two, draft three and now we're almost straight going to draft three.
Speaker A:And starting from that point, it's like having someone to just bounce your ideas off straight away and not being worried.
Speaker B:Yes, yes.
Speaker B:Sometimes it just replaces the search.
Speaker B:So things I would previously just look for on the Internet.
Speaker B:Now it is often faster to ask a null alignment.
Speaker B:Sometimes I ask the same question from different several different LLMs and they, they give me slightly different answers and it feels like, you know, I have consulted a few knowledgeable experts, but I'm getting some range of opinions.
Speaker B:So that, that, that's also a very nice feature of it.
Speaker A:Do you still use Google or.
Speaker B:I still use.
Speaker A:Search engines are available.
Speaker B:I still, I still use, use Google and search engines because yeah, sometimes I just want to have a look myself at the sources and often LLMs can hallucinate.
Speaker B:They're not always reliable.
Speaker B:So I think it's very important to verify what they're saying and sometimes when I want to definitely check what, what, what sources exist, what information is there and whether they haven't omitted anything because sometimes they give you the sources.
Speaker B:But It's a, it may be a biased selection and I want to, to to kind of have a range of, of of sources than I would do my own search because even when you.
Speaker A:Go on Google nowadays it sort of gives you like an AI summary at the top.
Speaker A:So I'm wondering well which LL are they using or how is it all pulled together?
Speaker B:I think they're probably using Gemini because Gemini is the Google's AI.
Speaker B:Yeah.
Speaker A:So yeah.
Speaker A:What other tools do you use?
Speaker B:Well I use yeah as I've mentioned my go to tools kind of all the time when I just want to talk to LLMs are Gemini, ChatGPT, Claude.
Speaker B:But I also sometimes use Perplexity and then there are tools like mindstudio which allow you to quickly create your own AI agents where you want to.
Speaker B:To put together some automations.
Speaker B:I also sometimes use Gro and yeah a few, a few other things.
Speaker B:I constantly play with something.
Speaker A:So yeah because new, yeah new job titles like Prompt Engineer are coming along and AI agents.
Speaker A:I hadn't even heard of that, you know, this time last year.
Speaker A:So it just goes show how fast the landscape's changing like the job landscape, the AI landscape, just, just everything because what do you.
Speaker A:Yeah.
Speaker A:What advice do you have for anyone that wants to sort of keep on top of it?
Speaker A:I know we said earlier what the start that there's too many newsletters but are there any that you recommend?
Speaker A:Like do you have sort of go to news sources or is it just whatever sort of pops up?
Speaker B:Well I, I don't think I have some go to new specific news sources because as I say I subscribe subscribe with to several newsletters but generally speaking I just go on LinkedIn and there are a few people I follow and I look at what they're writing.
Speaker B:I also regularly look at, you know, arXiv.org where people publish new academic papers.
Speaker B:So that's just to see what research papers have recently appeared.
Speaker B:I still look at Hacker News, although it's not a dedicated AI resource.
Speaker B:But this is where people discuss technical news.
Speaker B:Hugging Face is a very good source to learn about new models that are coming out.
Speaker B:So I'd say these are my top resources.
Speaker A:I suppose when people have published things then it is just out there in the public domain.
Speaker A:So yeah just like we can find it on those sort of websites then the AI is good to go and copy from that.
Speaker A:But do you think there's that?
Speaker A:I don't know.
Speaker A:Should we be sort of putting things in place to sort of say no, AI can't take from this?
Speaker A:This is our original thinking, or is it just kind of all in one big melting pot now?
Speaker B:You mean, how do we protect our own creativity from AI?
Speaker B:I don't think we can do that because it will either be something that is decided on the legislative level or it's not really practically possible because you can put a disclaimer, you can't use that word, but there is no guarantee that it won't be used.
Speaker B:So, yeah, I think at the moment, if you definitely don't want anything to be accessible by AI, it probably shouldn't be on the Internet.
Speaker B:Yeah, that's, by the way, a dilemma because there are some kind of texts I'm sometimes writing, which I see as something just creative for myself, and then I sometimes don't do certain tools because, well, I mean, AI tools on the web.
Speaker B:Because I think I would rather not see this text returned to me by AI one day before I even published it.
Speaker B:So probably the probability is low, but it's not zero.
Speaker B:So I think.
Speaker B:Yeah, when, of course, there is the exception is certain deployments.
Speaker B:The companies do them when they're dealing with sensitive information and they're creating corporate systems.
Speaker B:Then it's possible to deploy AI in such a way that any documents you are feeding it, any information you are giving it, will not be used for training.
Speaker B:But it's also possible to use AI locally if you're not a corporation.
Speaker B:But it requires a few kind of special tools and a certain setup.
Speaker B:But the majority of the users who are just going on the web and using AI tools on the web, it's not possible to guarantee that anything you are feeding it will remain confidential.
Speaker B:Not at this point, but this may make change, of course.
Speaker A:Well, especially if you've got AIs talking to other AIs.
Speaker A:Even if we trust each other.
Speaker A:Oh, no.
Speaker A:They might be listening.
Speaker B:Yeah, yeah, that's that.
Speaker B:That's that as well.
Speaker A:Oh, wow.
Speaker A:So have you got any advice for anyone that's listening that might think, you know, that AI isn't for them or they don't know where to get started?
Speaker A:You know what, what would you want people to take away from our conversation?
Speaker B:Well, I think that overall there are reasons to be optimistic about AI.
Speaker B:I think it gives all of us new opportunities, new capabilities.
Speaker B:And although there are things to figure out, there are certain risks overall.
Speaker B:Hopefully it's a force for the good and I think people should just embrace it.
Speaker B:So I think that my advice is if you are not yet using AI, start using it.
Speaker B:Just play with it.
Speaker B:See how it can be helpful.
Speaker B:Try to maybe automate some of the repetitive tasks that you're doing with AI.
Speaker B:And also really experiment with writing good detailed prompts because the quality of output differs very much.
Speaker B:Whether you are just giving them a short and not too detailed instruction versus when you are providing a detailed prompt and describing exactly what it is that you expect and what your criteria and what the good looks like, you get vastly better results.
Speaker B:And this is, I think, is not something that people fully realize yet.
Speaker B:So my advice would be to really go for it and experiment.
Speaker A:Thank you, Lyudmila.
Speaker A:So it's been so great to speak to you.
Speaker A:If people want to connect with you or ask you further questions, what's the best place to find you?
Speaker B:Probably the best place to find me is on LinkedIn.
Speaker B:So yeah, I would be very happy to connect and have a chat.
Speaker B:So please reach out and write to me and share your thoughts on AI.
Speaker B:I would love to hear how people are using AI, what challenges they see.
Speaker B:Any particular success stories as well would be great to hear.
Speaker B:So yeah, look forward to connecting with people.
Speaker A:Fantastic.
Speaker A:We'll put all the links in the show notes.
Speaker A:Well, Dr.
Speaker A:Lyudmila Lugovskaya, thank you for coming on Women with AI.
Speaker B:Thank you very much for having me.
Speaker B:Really happy to be here today.
Speaker A:Sam.
