[00:00:00] Peyton: Hello and welcome back to another CareerCon, uh, interview. Um, I'm Peyton, our web designer, developer. Um at hundo and today i'm today i'm here with um, Kenn Hello Kenn. How are you?
[00:00:23] Kenn Mayfield: I'm doing very well. Thank you peyton. And thank you for having me.
[00:00:26] Peyton: Not a problem at all Thank you for uh coming on today. Um, could you just introduce yourself and a little bit about who you are and what you do?
[00:00:34] Kenn Mayfield: Certainly. My name is Kenn Mayfield. The new name that I'm going under is XyrisXR Design, and I use the nickname XyrisKenn online. My background is in fine arts, but through time I've worked with more and more digital tools, both visually and then in programming iOS games and interactives. And now I'm combining AI with Unity through SingularityNet and my own projects.
[00:00:58] Peyton: Lovely stuff, uh, and I'll jump into the first, um, couple of questions. Um, firstly, could you provide, uh, any insights into the role that AI is currently playing, um, particularly in the workforce and how it's shaping various industries and career paths, especially for sort of young people?
[00:01:16] Kenn Mayfield: Certainly. So much has changed over the last year and a half, even insofar as AI is concerned.
When I was working on a project to develop an industrial metaverse with a company called Carbix, we'd started with avatars that would just use scripts to recite what they were saying from recorded actors. Then that turned into synthetic voices, and then that turned into conversation via chat GPT. And leading the charge for AI were these brilliant minds who were creating fashion.
And, uh, science fiction and other art on LinkedIn and other sources. So AI has really come in like a wave to change the possibilities of what can be done. And not only that, it's rapidly evolving to allow you to create things that can be explored in a more immersive way, instead of just 2D, we're now looking at 3D and capturing environments.
So how does this affect a career path? I think that The strongest situation, if I were to imagine myself in that situation, would be to continue to develop my personal skills, but to learn how AI can be used to augment that. Because what's happened is that AI has, as with previous inventions, like desktop publishing, has accelerated what's expected to be accomplished in a finite time of work.
But it's also risky, because these AIs were developed on data sets that were, you know, available on the web. So there are questions of copyright. So to understand how the process works, what it can offer you and augment your skills that you're developing personally, that are independent of AI, I think makes the most powerful combination and helps to future proof you, uh, as to what's coming.
[00:02:58] Peyton: Fantastic. And sort of leading on from that a little bit, um, how can young people prepare for the rise of AI, um, and machine learning? Are there any tools you would recommend?
[00:03:10] Kenn Mayfield: The most commonly used tools as I see as primarily a visual developer would be in terms of pure AI using Python and R is very common.
The Process of educating oneself as to how to create models is outside my purview, but I've learned enough to work with machine learning specialists who are very enthusiastic about what they do. So I ask questions like, what kind of technique will they use? Will they use this kind of neural net? And I rely on their advice.
So now AI has become such a mix that I think being able to keep current with what's being developed. Both in the chat GPT conversational side and then say the mid journey or stable diffusion visual side is helpful to understand how prompt works if you're an artist. If you're a programmer, then delve into, uh, these machine learning sources and programs to learn more.
[00:04:07] Peyton: Lovely stuff. And, um, in what ways could virtual work experience in AI benefit those students that are looking to sort of stay up to date with those newer trends, um, particularly for sort of younger professionals just starting out in their career? Um, to sort of help develop their understanding of AI technologies.
[00:04:26] Kenn Mayfield: We've had a really interesting experience and unfortunate, but also a powerful experience through the, uh, pandemic that started in 20 and then went to 2022. We discovered that decentralized work is really beneficial for people in family situations, for people who live outside of city centers and for people who work with disabilities, who have disabilities, uh, because suddenly they didn't need to dedicate so much of their day.
So I think that we should maintain that decentralized approach to work. In terms of AI development, a lot of it is happening online, of course, but if you can make your way to gatherings to meet people person to person, now in a safer context, uh, with regard to COVID, then that's important as well, because I think that we've come back around to needing to have some face to face contact, but of course there are risks.
So AI should support our decentralized Uh, work from home or work from outside the office initiatives, but there are other, um, opposing forces management wise that want to try and bring us back into the status quo. So, it's a tricky question to answer. Uh, there are two sides to it. Uh, there are enough resources online available to really augment your knowledge and keep current and LinkedIn a lot to build teams.
Um, but being able to meet people face to face, uh, at least once or twice a year will also be very positive. For your development,
[00:05:51] Peyton: I'd say that's probably something we use quite a lot at hundo when it comes to remote work and remote collaboration, um, using AI specifically in tools, software, um, I think has become even more increasingly valuable for us, um, especially at the rate that we sort of have new documents and things coming out and being able to find those summarize meetings, recordings, things like this.
It's really valuable, I think. Um, I suppose also, slightly sort of learning new things as well. Um, I have over the years sort of learned an awful lot from YouTube, and I've noticed YouTube trying to add some quite nice AI summary and topic understanding. AI sort of hints in there, um, which I think actually have potential to be quite handy Um, particularly on sciencey and mathy videos.
I've noticed, uh, if a topic comes up Um, there's some quite handy little auto generated tooltips that appear underneath to sort of explain those to users. And I think those would be quite valuable going forward. Um, for sort of educators and parents, um, what steps could be taken to effectively prepare students and young people for that increasing integration of AI into education, tools, society, and the workforce on a whole?
[00:07:10] Kenn Mayfield: Hmm. Well, that's a very interesting and complex question. I think that, uh, from an adult or parental point of view, what I perceive is that students today have much more kind of public visibility in terms of building a persona versus building your personality. So, a kind of a media literacy. Would be really valuable for parents to look into, so that they can teach their students or their children how to recognize, uh, AI generated content, how to recognize AI generated accounts or sentences and so on now on the educational side, uh, to be able to, if you're at a young age to introduce AI in terms of, um, Play environments in terms of things like Roblox, uh, supervised, uh, for older, older students.
I think one can begin to get into, uh, the sociological side. of how AI is affecting work and affecting our choices, both personally and on greater levels like nation state levels. For work, there's a lot of opportunity, both if you're code minded, to focus on the creation of models, to learn to identify what are useful data sets, how data sets need to be augmented.
Uh, to work well with machine learning models. There's a lot of complexity there. If you're of a visual, uh, bent and prefer to be an artist, then there are a lot of tools, uh, to perhaps start to model AI models on your own work. Um, if you're a dedicated visual artist, you might have a body of drawings or photography.
Those can be built to create your own bespoke, uh, machine learning model to produce new work in an AI context. And this applies for writing as well. And these get into important conversations about ownership of content. Uh, which is still up in the air, so one has to almost learn or teach it from a multifaceted perspective, what you can do, what you can learn, how it can be used, and how it can be perhaps misused as well, and this will place you in a stronger place for, for utilizing it safely in the future.
[00:09:20] Peyton: Yeah, absolutely. I think. Um, well, you sort of mentioned about, um, AI being used in increasingly gamified, um, contexts is quite important, especially to get lots of people using and learning about it. And not only how you can use AI as a sort of an output, but also how to create the actual models themselves, um, is increasingly important.
And then also I suppose, in educational context, um, I think, you know, Increasingly important is teaching students and learners to use things like chat gpt as a tool as a learning tool and not an outright essay replacement, um, and use it as another sort of parallel to Google almost, um, to answer queries and sort of stretch your knowledge a little bit.
Um, at the end, they touched on the ethical considerations that might come along with AI. Could you just sort of elaborate on that a little bit? Um, Sort of what considerations should educators, parents, and even policymakers in government and legal fields take into account when integrating AI technologies into educational settings, particularly for those sort of ethical or moral concerns.
[00:10:34] Kenn Mayfield: Interesting question. I come across those as I'm developing proposals to build AI, which are also combined with 3D. Thank you. So the problem is that, uh, the AI revolution as we see it now, although it's built on decades of research has suddenly spilled into the public domain. In terms of its use very quickly.
So the problem with the datasets themselves are questions of who owns the original training data. Uh, as a user, the questions become how much am I relying on the presentation of this tool to create a, a, uh, career for myself? Uh, now what's happened in the past with new technologies? Even in music has been a great wave of use where all you see is the use of that tool, say, synthesizers or automatic performance, and then that becomes subsumed into real music performance in terms of an expression of the artist rather than the machine.
So, with AI, the question becomes, how does this augment. Your communication, whether you're a scientific writer, or you want to study philosophy or you're in economics or programming, um, the ethical issues will certainly come into it. There are also questions of privacy. To my knowledge now the EU has the strongest privacy laws with regard to AI.
Now, if I were a student using AI, I could offer advice from how I use it for proposal writing. I find when I'm writing, I will get into kind of a habit of how I speak or using many words, because I grew up with books, um, 19th century and 20th century books with lots of periods and commas and semicolons.
So when I ask chatGPT, please rewrite this for clarity and use this It gives me interesting feedback, but then I should process that back through my own writing and just take what seems to make sense to try and maintain my voice because the other flip side of AI is that the styles tend to be recognisable from one user to another.
[00:12:38] Peyton: I think for me. Something quite important is also the accuracy and the copyright of them. So if there's potential for models to produce potentially inaccurate, harmful, or otherwise just, um, amusingly false information, um, and another quite important thing I read recently was, um, the, the very fine line in, uh, AI legislation sort of surrounding, um, blockading.
Large companies in almost garden walling. Um, so something to be quite careful of is, is not, you don't want, um, a collection of companies to build some really powerful AI models and then for legislation to, it tends to lag behind by a few years to then catch up and essentially block any more innovation in that space, um, which in effect almost creates this, this, this.
monopoly amongst a handful of companies. Um, and there's a fine line there between taking swift action to keep people safe and, um, to monitor big, large companies, but also not taking such strong action that you only limit that possible innovation to a select number of people. Um, Yeah, those are the two sort of things I would add there.
Um, I've got one final sort of curveball question for you Kenn, or two, I suppose. Um, could you explain sort of, uh, one thing that scares you the most about the future of AI and one thing that excites you the most about the future of AI?
[00:14:18] Kenn Mayfield: I might be able to answer those two questions with the same answer.
What scares me about the application of AI, first of all, is the invention of popular opinion. That if one is spending a lot of time on social networks, it's easy just through advertising mechanism and scripting to create accounts to post opinions and that can have a direct effect on just the gestalt or how people think.
In general, so we need to be aware of that. I'm also concerned about, um, people say neurotype, autistic, not having a place at the table in terms of how the web develops from AI in the past accessibility has been a bit under supported and models currently do exhibit some bias, both in terms of culture and in terms of neurotype.
So I'm concerned that the web 3 would develop a full speed. Along similar patterns to web two, but right now we have this great window to be able to affect it. So I'm excited to be able to work with companies like singularity nets and deep funding, which provide people like me, a window into learning how to develop our own artificial intelligence and also our own future artificial general intelligence.
There's actually a tool set that they've recently released to gradually work towards that. So on one hand, I'm concerned about the misuse Of AI as with any other tool, but on the other hand, I'm also excited of how it presents a pathway to the future for people like me, who have been a little bit, um, um, prevented from the full benefits of the mainstream to date.
Yeah, we can now build that ourselves.
[00:15:58] Peyton: There's such wide range of use cases can be on the huge spectrum of incredibly helpful to. Um, uh, maybe the other end is hacking, spamming, et cetera. Um, and obviously you sort of get a bit of both. Um, something, I think that's probably something that excites me a little bit as well, is, is just the, uh, excitement of seeing where the ceiling is for AI.
Are we plateauing at somewhere soon? Um, or have we got a long way to go? Yeah. And I suppose something to add to that would be a sort of related. Um, thing that scares me a little bit is, will there be a point that we hit, um, such, uh, such a high mass of AI generated content, um, particularly on the web that we've almost drowned ourselves out and we're no longer able to progress our own models.
And then we've sort of built ourselves into this zone, uh, into our own plateau. So, um, so I'm intrigued to see. What the future of that holds for us. Well, thank you very much, Kenn. Uh, I've really enjoyed our conversation. Um, it a pleasure. If people would like to connect with you or find a little bit more about what you do, where is best for people to find you?
[00:17:16] Kenn Mayfield: Okay. I can be found on our website, which is xyris.ca x y r i s . c a if you do a search for XyrisKenn X-Y-R-I-S-K-E-N-N on the web. You a lot of my. Other work will appear and through Singularity net deep funding and of course hundo.
[00:17:35] Peyton: Well, lovely. Thank you very much for joining us today
[00:17:38] Kenn Mayfield: It's always awesome. Thanks for having me!
