AI Foundations Program Workshop #1: Introduction to AI

0:28 for e

1:58 e e

2:46 shonas I think we can start now everybody is joined uh here in zoom and also we are live streaming in YouTube so everyone is seeing the same things so shonas the stage is yours uh perfect uh so thank you very much uh thank you for this opportunity I'm super thrilled uh to join um this uh women go Tech program um and um hopefully I can see um messages in chat um so whoever has joined I see that it's hundreds and hundreds of people which is super exciting can you please tell me which country uh did you uh join from um to

3:34 watch this um lecture about the AI so if it's possible uh to write in in in chat I'll be uh happy to hear that if not um I'll I'll kick off with uh with um our um lecture so um obviously U we are talking today about AI uh something that we hear daily and something that is probably shifting our day-to-day uh usual um working hours um how we perceive everyday tasks um so um in in today's lecture um I will try to give you an introduction uh about the AI so we have the same basis of understanding of

4:20 different AI tools and what AI is and the AI history and about the Everyday Use um of AI as well so to be uh to be more concrete um we'll have um um couple of um um we'll talk about a couple of things today first of all we'll talk a little bit about um like an introduction to AI then we'll talk about tools for everyday use tips and tricks with AI and impact U and ethics of AI obviously this is uh one of many uh lectures that you will have so in this lecture I'm not going to cover everything um so first of all uh we will do a a quick AI test uh we will talk

5:07 about why is it important to talk about um Ai and digitization in general uh we will talk about Origins and evolution of AI where do we use it daily and what is AI in simple terms um so I can see a couple of people already telling me where they're from so I see people from Czech Republic lethania Ukraine uh Lithuania Berlin Estonia Slovenia ltia many lithuanians someone from Greece Poland ltia Georgia um ha to tibilisi Hungary Romania uh so many Bulgaria Italy uh Poland Estonia Romania again uh so many of you um it's super exciting so hi to

5:54 everyone and and each one of you and greetings from villus I see people from Malta as well um from Mexico as well partly um that is crazy and it's it's super exciting so we can have a a joint session about um AI from all over the world um so if you have things that you want to share with the group obviously you can use chat and I'll and try to do my best to um keep an eye on chat from time to time so um love to you all from villus from lethania and let's kick off with our lecture so what I'll show you now um is a series of um pictures and one of these pictures is created by AI

6:40 another picture is a real picture usually a photography so your uh task now is basically to look through the series of these pictures and tell me which one was created by AI so left u a or right B it is always going to be left a right B so the question is which one is created by Ai and we can write it in chat so we already have a polarized answers but uh most of you are telling me um a so this conrete picture um picture a which is on the left was created by Microsoft Bing uh and picture B is a real waterfall uh from Iceland um so let's kick off with um

7:28 another picture so picture picture a on the left and picture B on the right which one do you think here is created by AI I can see that uh many many B's some A's as well uh so again we have uh um different opinions but most of you I see are giving an answer B so picture B was created by m journey and picture a is a real Gondola in Venice so again you probably now we only had like a couple of pictures but you already probably can tell um some ideas at least uh which picture is made by AI maybe some Clues maybe it's something with the color maybe it's something to do with blur maybe it's something to do with you know

8:14 each Gondola being really similar to one another maybe some details are missing so you can already be telling me uh which one is created by AI so another example a on the left or B on the right which one is created by AI um this one again is a bit more mixed but I see more BS so sometimes when we create pictures with AI we get like a six finger a thired leg uh but now with generative AI tools we uh usually get you know better and better results each day so um this is um basically created with a quick prompt U by mid journey and picture B is mid Journey picture a is a real photo again uh another sequence

9:01 which one do you think is um generated by AI tool I see many of you uh some of you are saying both um also also good answer uh so A or B okay so many A's and some B's as well so I have a joke on this so um a guy super happy with a salad bowl obviously it's cre by AI it's not a real picture so picture a is created by moury picture B is a real photo of a woman enjoying a salad with a nice view yeah exactly vaa men do not eat salad at least they're not happy when doing

9:47 so so another picture uh which one is AI uh left a or right B I can see many many A's some B's as well so again we can probably tell that uh you know picture a is quite perfect uh so picture a to me looks like what you see on the internet before you buy and picture B to me looks like what arrives after you order something online uh so picture a was created by m journey and picture B is a real uh picture in the aquarium um and the Coliseum a on the left or B on the right which

10:33 one is AI okay so I see most of you already unlocked the answer which is uh you know sometimes a bit bit of a shadow maybe is casting in a in a weird way sometimes the picture is just too perfect so uh picture B was created by moury one of the most popular uh generative tools uh for creating visuals and picture a is a real picture of the Coliseum um another example um so we have some fireworks a on the left B on the right which one is AI we have different opinions but um I'm super excited that

11:22 that um I kind of uh managed uh to um um to kind of uh uh make my point of view a little bit so picture a was created by m and picture B is a real photo so I've been showing you series of pictures where like more colorful uh pictures were usually Ai and this one is uh um different this one is is is basically the photo camera opened up the aperture for a bit longer and then it closed uh so that's why it's so colorful and picture a was created by by moury um so uh sometimes you know you can prompt a picture which uh which

12:08 looks more natural U which looks less impressive and in that way uh you know create a like a a misinformative u um picture and last one um summer holiday in Lithuania or in Poland or in Czech Republic or in any Central and Eastern Europe a on the left or B on the right which one uh is the real picture again mixed uh mixed opinions so what we have here is picture B was made with moury on the right and picture a is real palm trees in the tropics so I know this is um um like a a

12:57 really uh quick uh test that we are doing now uh but when we talk about you know serious things like uh trying to understand where whether it is a disinformation attack whether it's a real information uh finding that information on the internet and finding reliable sources so we are going to have more and more of these issues and we have to keep really mindful um about what we see on the internet because this is just me kind of playing with AI or not AI um but in real life we we really have to pay attention on what information um uh we base our our truths upon um so I want to quickly introduce myself as well uh we started with a with a quick

13:44 and fun test but uh my name is shaunas shuna digitus uh so we can connect on LinkedIn uh I share from time to time U some interesting facts about AI um and and some not just like quick facts but some deep analysis as well um obviously usually I do it in Lithuanian but from time to time I try to do it in English as well so I'd be excited to um connect with you guys um but the information that I'm going to share today most of it is coming from me as being a tech Enthusiast so I have a weekly technology radio show here in liania where I talk about any Tech news that come up this week um I'm also uh a board member of

14:30 the ech association we are basically an umbrella of Technology initiatives that are trying to create breakthroughs in education so any Tech um Tech startup in education is basically Under the Umbrella of the ech um in which radio it's in jadas uh in in informational radio show here in Lithuania and I also do uh these sort of lectures and talks and and in-depth trainings uh for organizations um so what I just show you is just some fun pictures but um there was um um a bit of a deeper te test created by AI 21 and um what this test uh actually did was

15:18 basically um there were one one and a half million people um in this test and we had more than 10 million chat windows so basically in these 10 million and more chat Windows um we had um people who were joined um with other people or AI for a two-minute conversation so basically let's say you are a participant and you join in this um AI 21 test and you get two minutes uh with AI or a person then you get another two minutes with a person and you have to after this two-minute conversation you have to tell whether you were talking to AI or you were talking to a real person so an interesting thing occurred that

16:04 60% success rate was uh basically people telling that they talk to AI or a person so 50/50 is a guesswork uh so when we basically take a guess I don't know so it's a 50-50 shot so only 60% success rate of the people talking to AI um is um I mean it's quite um it's quite crazy so um I mean we have to be be mindful not only uh with you know uh looking at uh you know pictures that are created by generative AI tools but we have to be really mindful when we look at comments that are being created by uh bot farms uh that we are looking at I don't know any other conversations that are created in other social media platforms like

16:51 YouTube and so on so um each text each person uh that we find in digital world we we have to really be mindful about whether it's real whether it's truth or not because um in these uh disinformation times misinformation times um we have to be mindful so uh keep your eye on this um it's um not as easy and it's not going to be easier in the future as well so we have to be mindful at least and these are just um a couple of examples these are free conversations uh from this AI 21 test and you can just you know um try and see whether you can uh you know see which one in these three different

17:36 conversations is AI and which one is a person so I'm not asking you uh to uh actually uh try and tell me but just basically just see whether you can tell which one is AI and which one is a real person so you know we can take one of these scripts so a short poetry please yo Rose is red blue happy birthday wish I happy Birthday homie thanks let's drink what haha sounds like a plan to me what are we drinking coffee butter wish prco so you know these are U many different slangs uh many different sayings but uh um it's really difficult to tell which one is which um so um to make it easier for you AI is on the left

18:23 in each of these conversations so to me AI is on the left uh so all of these black uh um uh are Ai and the green ones are human so these are three different uh conversations um and again it's quite difficult to tell which one is which so again why is it important to talk about Ai and to talk about digitization in general why are we here so many of us uh and why is it important so um I've got some really interesting stats uh from EU so one of the main reasons when we talk about uh you know grown-ups and we talk about their digital skills so we have a bit of an issue so whenever it's a a

19:11 person who is a bit older obviously we have a bigger issue so at least here in lania the teacher digital skills are quite you know quite a big on of an issue and it's it's an elephant in the room and I'm pretty sure it's in in other countries as well because you know little kids now have iPads and they are using AI tools every day and you know um school um teachers they are basically just starting to use AI tools and uh so you can probably see where your country rates but when we talk about digital skills uh so we we not we're not just talking about you know how to use a laptop how it works um but we're we're more like more and more talking about

19:56 how can I find um information on the internet how can I make sure that that information is correct how can I use these new AI tools to help me to be more efficient to help me to be more productive so these are more and more of an issues uh that come up under this umbrella that we call digital skills and it it expands uh every every day because more and more tools are coming up and more new technologies emerge like AI that we talk talk about Ai and why we talk about digitization is that we have um an estimated number of displaced jobs due to Automation and this is by 2030 so we

20:43 have different countries here and we have millions and millions of people who are going to be displaced because of you know new AI tools are coming in not only generative AI because new technolog is coming in and is it's automating our day-to-day lives so you know when we talk about automation um one of the challenges that we have to understand is that uh not only uh you know simple things like cashier job is being automated where you don't need a person to scan your food when you buy it where you can order it online or you can just go through that you know quick cashier which is automated but now we are talking about marketing Specialists now we're talking about illustrators who used to you know uh create some you know

21:30 super cool visuals now we're talking about uh sometimes partly web designers now we're talking about even um let's say junior programmers who are in this situation like we still need more uh programmers than we have now but we need mid or senior programmers because where would you spend your money uh so imagine yourself being in like ahead of some you know Corporation and you have a team so where would you spend do you spend € 20 on chat gbt where you send your code uh it's you know it has some sort of mistakes so that chbt corrects that mistakes in you know 30 seconds and you

22:16 get the perfect code you paste it and it works or you spend like a couple of thousand for junior programmer and that programmer is basically you know coming with his own uh or her own ideas and then you have to give the feedback and then you have to do this and that so it is it is an issue um and um it it's something to think about um it is still um something that is worth talk about something that's worth discussing something that's worth to be a part of um but we really have to think about you know other competences that we have to come with whatever that we're learning now um so one of the um good sides of being in a tech industry you can

23:02 probably tell it yourselves as well so the um it sector is still ever growing so this is one statistic from lethania that I wanted to show so if you have anything plus de that means that your salary is at least um half as big from the general salary that we get here in lethania so let's say you don't even want to be like a programmer but you know some tech language and then you become like a project manager the tech company so your salary is at least 50% higher than the normal average rate of salaries here in liania at least so anything Tech is actually uh quite important as well um and um another interesting statistic that I want to share so more than half

23:49 of kids that we have in primary schools today are going to work jobs that don't exist yet so this is from World economic forum um so I want this to really sync in so 65% of children entering primary schools today who are actually in primary schools today are going to work jobs that don't exist yet so we are basically training kids uh with whatever the academic knowledge that we do um and when they finish school more than half of them are going to work jobs that don't exist yet and this is where we can you know think about our an analogy so from my perspective I didn't know that I could be like an Esports person so um Counter Strike uh MVP who is a lethania

24:36 now um you know he's a an Esports Enthusiast um maybe um like an influencer maybe a YouTube streamer maybe someone else that you know when we were at school we had no idea that we can be one of those and now you know we can actually have uh more of that which is quite exciting and uh it's quite exciting for those kids who U are going to enter uh the whatever uh work field that they will in the upcoming time um so the last statistic that I want to share that um 150 million new tech jobs are going to emerge within five years so um this is again World economic Forum um

25:21 and um it's important to understand that you know technology is the way to go uh even if you think about you know a simple business which is you know let's say I am selling you know I'm a b keeper and I'm selling honey so if I want to sell my honey obviously I'll have to you know have bees I'll have to have like be hives and you know I'll have to care for them um you know I'll have to care for them during winter but if I want to sell anything that I make I probably will need to have a social media I'm might need to have if I have like a big Barm then I might need to work on my SEO uh on my Google ads I might need to have like my um website to to share more um

26:11 where you know people could find me um a bit more easy so any job that you can think of it will have a tech part even if it is nothing to nothing to do with uh like technology let's say you want to do a retreat Yoga Retreat in Thailand and you want to do it without technology but how are you going to get those people into that course it's going to be via socials uh or you know some sort of you're going to enroll them through some sort of uh CRM some sort of management system that you work basically via your laptop so any business is Tech and new tech jobs are going to emerge in the upcoming couple of years and they're going to be at technology jobs so um moving to um Ai and talking

27:00 about Ai and history so I want to um establish uh some sort of uh uh breaking points at Ai and I want to do it through uh 1950s to now but I also want to give you a start a kickoff with Gary Kasparov uh so he is one of the greats this is not real Gary Kasparov by the way this is my attempt of creating Gary Kasparov with me obviously it can create a real gasparo face um and U obviously you can tell that this is generated uh with some tool because the chest pieces are not right uh so um you know if you are mindful you can tell um and um what I wanted to tell

27:49 you about this um is that in 1997 um de blue U Beed uh beat uh G part of at chess so for my generation that was you know sort of wow understanding that AI can actually beat like the you know one of the smartest guys alive in that sort of his narrow field of Chess um so 1997 2.5 3.5 IBM won against Gary Kasparov and everyone sort of was amazed and thought that AI is now going to take over and it's just going to change the world that we see um but um everyone forgets that 1996 Gary Kasparov won

28:35 against the blue and everyone also forgets that IBM actually bought a team of guys who were creating this machine learning algorithm called dblue and these were fungu Thomas anantaram and maray Campbell so these were you know the the core of the team there were more people involved obviously um but um we also have to understand that this is uh you know 64 squares on the board and when we talk about um you know this this this big picture of of AI is going to change the world we have to understand that AI can change the world of a narrow field so chess as we know it is different than what it used to be chess pre 1997 so if you play chess like me

29:23 I'm a chess big chess geek um so you know now now because uh before 1997 you know if you were like a a world champion let's say you had your own team you had to prepare for chess games and then you go and then you compete and you analyze games and then you go and compete again so what happens now is basically you have a pocket phone which is stronger than any imaginable team that you can have of all the MVPs in the world and you have the uh data of all of the people that played all of the games games throughout the history so and and and it's all you know it all fits in your phone in your simple iPhone and then obviously there's machines like uh fish

30:10 stock of a machine playing against another machine and these machines play in a different manner than the humans play and you can tell and the people that play chess now they play differently as well because now technology emerged and merge with chess and now it's a different game alog together so AI can change narrow fields and AI can dominate narrow Fields as well but when thinking about AI changing the world around us so um I wouldn't go that far and we can uh talk further a little bit um we can talk more about this a little bit further so a short history um about AI so what I want to tell you here um is um you know some

30:58 um uh in each decade some main points how AI evolved and what we had within this um AI field uh go game Ellie by the way it's a it's a Chinese chess in in another way um and I'll talk a little bit about it as well because in 2010 if I'm not mistaken um alphao uh which was by Google um they beat the uh Chinese chess or uh go chess um um you know master so it's it's another it's another story as well but it's another breakthrough that we have so uh let's kick off with 1950s let's let's understand the the AI where it started where it kicked off so in

31:44 1950s we had the the touring test which was obviously proposed by Alan touring um in 1950s to determine if a machine can exhibit intelligent behavior um um in distinguishable from that of a human so basically the the test was created to distinguish whether it's a human or a machine um and um in 1950s we had the terminology of AI created um which was introduced by John McCarthy in 1956 uh it was in in in in Dartmouth conference by the way so 1950s we have touring uh test and we have the terminology of Ai and we have an MIT

32:29 University which basically you know started talking about AI started creating AI courses and so on um so what 1960s brought us was the birth um of Eliza um and Eliza was basically an early AI chatbot uh so it was basically created uh on NLP uh on neural linguistic programming uh so you basically can talk to Isa and it or she would respond by saying well how do you feel about that what do you think about that like a like a psychologist would um so that was like a it wasn't like a large language model that we are used to today but it was the

33:16 beginning of it all um where you can have a chat with a computer let's say um in that way um 1970s we have um uh big data used in Pharmacy so we have ai uh we started using AI in Pharmacy um and it was basically diagnosed um it was basically designed to diagnose bacterial infections and um recommend like antibiotics to use so using that data you diagnose and then you recommend uh like a um antibiotics uh course for that person by knowing how strong uh you know

34:02 is that bacterial infection so by using these um big data and algorithms which which is basically AI um so what 1980s brought us is AI winter and what AI winter means is basically when uh we think that AI is going to change the world then we are kind kind of like going uh you know up uh and everyone is excited so so obviously Investments are really high in AI but when we reach that point where we understand that our expectations are not met then AI winter starts where uh investment decreases people's attention decreases we are still not you know we're not having any flying cars at

34:47 least not in lethania tell me if if if you know there's many countries here if you have flying cars then I'll be excited to come to visit your country as well um so we obviously you know still need to work uh and sometimes even more so expectations are not met and then you know it kind of uh goes a bit down and then that's why we call it AI winter and that's what 1980s was after you know um Liza after terminology of AI after using AI in uh in medicine um uh Tesla's gallor yeah aneska sure um so 1980s AI winter but we also had neural networks which is basically like U I will tell you more about you know

35:35 how to understand AI uh but neural networks is the basis of it all uh so we had some really big breakthroughs in neural networks in 1980s and the 1990s we obviously had a d blue which I already talked about um and uh we had machine learning breakthroughs so we have neural networks then a bit more complex AI is machine learning so so let's say neural networks is connecting data with this little synapses and uh machine learning is having uh more of these knots of information that are joined with neural networks and then 2000s brought us more for um breakthroughs in deep learning so it's neural networks uh machine learning deep learning which is you know more complex

36:23 knots uh of information that are joined with uh other neural networks um and again um in 2000s we already start seeing U AI for just simple consumers uh and we'll talk about our daily usage of AI and sometimes we don't even know it but we use AI not just in generative AI tools but just in our day-to-day lives that we sometimes don't know um so 2010 someone already mentioned in chat there was also alphago which was created by Google and um it was basically uh you know the alphao champion was beat by uh um uh goess Champion was beat by alphago uh

37:09 Google project and 2020 uh 2020 is where we uh where we live now um and it's I mean we are at those you know hype AI uh world we basically are super excited about it we talk about it every day we hear about it every day and hopefully we use it from day to day because we can really be more efficient and more productive when using different AI tools and hopefully my lecture my next lecture and uh a next lecturer after me is going to help you to uh work out you know in this you know Jungle of AI and use you know tools that are easy to use that are simple to use and are safe to use uh for you know Everyday

37:55 Use um so it's um yeah pretty pretty exciting times so um where do we use it daily uh so these are just um a couple of examples where we use AI daily um but uh we can obviously um you know let's say let's say we're meeting live here so all of us and then you know people watching on YouTube not just in Via Zoom people watching it later whatever but so it's you know thousand people or even more uh so you know we are at this event where I give this lecture and then after that we call Uber or bolt or whatever is more popular in your country in lania Bolt is the the go-to for some reason um so if we uh

38:41 call bolt then you know whoever calls B the last pays the you know highest price because machine learning algorithms work that way uh because they know like let's say you know there's a a peak of interest in this so that means you know the price go goes up so it's it's simple um it's simple tool um I mean not simple tool but it's it's it's simple algorithms in a way that uh you know work so we have um these different uh prices we also have GPS navigation so let's say you know after this lecture we are driving somewhere and let's say it's a peak hours uh so it's lots of traffic so whoever um you know uses GPS they send the information to satellites

39:28 satellites send the information back and then we analyze that information and then we know how many cars are in which route and then you know that GPS sort of navigates you through the traffic and gives you the best route usually what happens sometimes not but usually it drives us through the quickest route so it's also AI another daily example of AI is basically our spam filter so um a spam filter um is a again if we wouldn't have ai uh in our spam filter we would get all of these African prince messages of inheriting millions of euros and dollars and whatever and we would need to work at all anymore so uh this

40:16 spam filter actually filters out the information which is usually fishing attacks which is um us just just the usual spam um so then we have you know all of of the work emails all of the friend emails that are actually making sense to us so spam filter is also AI face ID so we can open our phones just by looking at them that's also AI C voice commands or chat gbt voice commands that's also AI Netflix recommendations so if you click on something that's interesting to you that day next day you might find Netflix recommending you the same thing or something similar something that you also might like which is is also AI um again predictive text uh on on your

41:02 phone that's also AI in a very small scale but it's also Ai and then content recommendations on social media that's also AI so even if we're not using generative AI tools we're using AI daily nonetheless if you have a phone now GPS email whatever laptop you're probably using AI every day so um this is a bit U it's it's oversimplified version of uh what AI is but I want to establish um the the basics of you know how we understand AI so uh so bear with me here I think it's quite important just to

41:47 understand Concepts so AI is systems or machines that mimic human intelligence to perform tasks so that's quite simple so can improve themselves based on the information they collect and AI typically requires data input to learn and make decisions often doing so faster or more accurately than humans can so that's AI in general systems or machines that mimic human intelligence to perform tasks so what is machine learning then so let's think of AI as an umbrella so under this umbrella we have uh neural networks deep learning and machine learning so machine learning is a subset of AI obviously um so systems or machines that

42:33 mimic human intelligence to perform tasks obviously can improve themselves based on the information they collect um again it's quite um simple as that um deep learning is the subset of machine learning so it uses neural networks with many layers that's why it's called Deep learning so many layers of machine learning is called Deep learning and it's particularly good with speech and image recognition so let's say well there's many there's many different examples but you can teach these deep learning models to you know understand you know which picture is an Apple so you give 10 different apples and then you give something else which is red and then it will say well this is Apple this is Apple this is just something red so

43:20 that's you know in a nutshell how it works and neural networks is a subset of AI inspired by human brain and it's composed of layers of nodes or neurons that process data so again with more and more data like all of AI it can improve um so you know this is you know some general terms of of AI and you will hear more and more of machine learning of deep learning of neural networks um as well um so um AI is the overall field machine learning is a way of achieving AI by letting machines learn from that data deep learning is a specialized type of machine learning with complex neural

44:06 networks and neural networks are the backbone of deep learning models so um this is in a very simple terms I see also people super uh um active in the chat uh so I'm really happy to see that and sharing some more information so you know I know that women go Tech Community is uh you know quite big and I'm super proud to be a part of it so you know obviously you have that resource of working together so if you have some you know interesting things that you want to share uh you have all of these you know other amazing people within this community so uh so please do please share um okay so we go to the second part of U our lecture um after you know

44:54 today's lecture you will have another lecture not with me but the third lecture is going to be with me again and I'll talk more about you know uh productivity tools so I'll go more in depth about how to use them you know what tools to use and and so on so today we are kind of you know creating that basic understanding of of of tools of of AI of history of AI and so on um so it would be really interesting for me to hear you know uh what is your go-to uh tool for day-to-day use if you use narrative AI tools so you can write it in chat as well but from what I know I checked a questionnaire uh or like a little research that women go Tech did so I know that most of you use chbt but I'd still be really interested to see

45:40 what other tools you might use or how many of you use chat GPT mostly so if you want to share please do in in the chat so we have perplexity Gemini uh Chad GPT Gemini and Chad GPT mostly Chad GPT quite quite a few lots and lots of Chad gbts and Geminis as well some more uh tools so all free uh already someone said so there's quite a few um and you know if you someone don't use yet which is also uh which is also okay so um if you want to share you know some tool that you might think is you know the best for you or you recommend it uh to anyone else then again use this chat use this community

46:27 to spread your ideas as well so uh from my perspective um chbt um is the go-to because I just enjoy using that you know it's it's um it's one of the Pioneers in this field and and and for me it's it's really easy to use the user experience is quite simple to me the user interface is also very simple to me so I use Chad gbt mostly but uh they're not paying me to advocate for Chad gbt I'm just a keen user and I enjoy using it uh daily and and it just helps me um to to achieve my means um another uh you know um quite popular to is Google's tool is Google's Gemini so each of these tools have like

47:14 a partly free version where you can use it partly for free but then again if you want to use it more in depth or if you want to use it more often then you have to basically pay for it more so that's how it works uh usually so let's say chbt you know if I use it for simple you know my you know daily tasks if I overuse it then I might not be able to uh uh use it um anymore because there's a limited uh amount of times that I can ask something for chat GPT um so with renewed version you can use AI assistance uh which is basically you know these gpts that are already created or you can create your own but with a free version you cannot create your own your own GPT or you can't use um um Dolly for example so you

48:03 can create text but if you want to create a visual then you have to find some other uh free uh generative AI tool so it's always um it's always an issue there uh so again Gemini Claud anthropic which is now mostly in us but in some Western countries we have uh more of it now as well and it's steadily but coming to Europe as well now it's I think in France and Germany um so if we talk about visuals uh so we have um we have uh quite um a few here um Dolly is obviously uh you know used with chat gbt mid Journey um is used um via Discord and Leonardo AI

48:50 is a partly free tool that I recommend to use and I'll talk a little bit about it in my next lecture but it's basically a tool where you can have daily credits that are renewed each day so let's say if you want to use a just a little just from time to time just for fun then it's you know very easy to use Leonardi and you're not sure if you want to buy a subscription of let's say some sort of other tool um so it's it's it's a nice alternative to have so dolly is used by um um chat gbt mid Journey via Discord um and Leonardo AI is basically um just another platform another tool that I recommend that you can be using uh for uh for free so um I

49:40 want to quickly like um show you U the user experience of uh you know these three different tools just so you have the idea of how it looks like how different it is if you haven't tried them yourselves so um just bear with me will try and share uh Dolly first My chaj BT and I'll share uh the screen again and show you uh M Journey um as well so just bear with me okay so this is my chat GPT um and whenever you use uh Dolly uh which is via chat gbt so you have it

50:25 here in the corner where you find your other gpts so this is your general user experience when you use chat GPT if you go uh to U chat GPT um uh to you know to find other gpts you will find you know here you can find uh Dolly and you can start a simple conversation like you would using simply a simple you know chat GPT um so uh let's say I want some sort of visual of a guy giving a lecture on AI for an amazing women go Tech Team all over Europe uh online let's see you know how this uh

51:13 let's see how this works so what it actually does now um so uh for some reason it's just giving me a text but let's let's try it again we'll try and see if uh creates a visual with its own uh created um uh prompt so as you can tell um you know what Dolly does well so it kind of rewrites your prompt to create a better prompt if you want um so um it's pretty uh it's pretty cool so here it is um it's uh it's what it kind of imagines of uh you know AI my lecture via Zoom uh looks like so this

52:00 is a a simple way of using Dolly on chat GPT if you ever think of uh you know using paid version and then using some sort of generative AI tool of creating visuals but I like it I see U it as a simple tool to use and um I sometimes you know recommend it to um other people as well so I quickly want to show you how mid works because you've seen um you saw a lot of M Journey pictures and um I kind of faked you into believing that one or the ever was created by uh um um um by a generative AI tool so I want to show you which tool was used uh to uh try and and do that so this is U um this

52:49 is a mid joury tool so how it works it's basically in Discord so Discord is I'm not sure you know how old every one of you are if you want to write in chat then please do but uh in my generation we used to have MK which is basically like a if you had mirk let me know this is uh it used to be like a place to talk um so we used to have like different chats and it was like you know in the 90s um and it was a tool where everyone talked to everyone and let's say like I was interested in skateboarding so I had like skateboarding claya skateboarding vus and then you kind of that's how you met people in those days that you didn't know that's how you kind of went to places because you just you know you didn't know people so uh so that's

53:34 that's how we used it and so it's very similar this is a place where you can chat to other people um but you can also have like a mid Journey bot so mid journey is integrated in this tool and you can simply write imagine um and then write your whatever prompt it is so um I can I can actually copy and paste this the same prompt and see you know what result will I get um uh everyone had milk by the way which is pretty cool I'm not sure if it was only like oh no I see like not only leanians had milk which was like probably the best tool that we had those days um was like with Windows 95 and uh so that was uh that was awesome um milk

54:21 was basically like a place where you used to talk to people so um it's U online obviously okay so you know how different M journey is uh um so it's um it's a bit different um we can tell uh you know already that we have you know four pictures instead of uh instead of two pictures um and uh you can you know choose which picture you like let's say I like uh uh picture number one more so I upscale picture number one and then you know I can work with this picture a little bit more but um you can see like creepy faces in the background but U women go te also here almost um but um it was before Facebook

55:09 uh 100% Mir was before Facebook so this is how mour looks like um it's it has many different you know things that you can do you can edit you can you know change things and it's mostly used for you know creating you know this inspiring crazy visual ual um Dolly is a bit more direct um but you know two different tools but very simple results actually uh by both of them so I quickly wanted to show you that and then last but not least uh uh I want to show you um this free tool that I recommend um which is basically if you just simply want to uh want to try it out so it's called Leonardo AI I just want you to see the user interface so you know it's

55:56 the safe website that I recommend or not something crazy and you can basically uh you know put your email via Gmail and then just quickly get started and you you can create these you know amazing visuals here for free and each day you have um 150 credits if you want to do more then you can do more these are just some random uh pictures that I created with uh uh School teachers that I help them with their um digital competences so it's a it's a it's a you know way partly free tool but I really recommend it um and um yeah it's um it's quite easy to use as well as you can see you can choose you know different image Dimensions different sizes and then obviously it works best uh with uh uh

56:45 English text so if you want to work with your language um it probably will not work um but a good thing is that um we can use um you know chbt to write our prompts uh we can use chbt to translate uh and we can use it for anything else that will basically help us to get the best results so um if you're stuck you know chat gbt uh that's my uh simple answer to many of my uh daily um issues that I have um so let's get back uh to our lecture I just wanted to quickly show you uh you know these tools for for creating visuals how they work how they look like if you ever thought of it um

57:31 um and quick answer to Sandra how much credits for one video uh for a picture it's 25ish credits for a video video it might be a little bit more but uh we need to double check on that okay so um uh since uh um since next uh lecture is going to be more about prompting more about creating effective prompt I will only give you like a quick Basics so you will have a general understanding of how prompting works and what is the you know from my perspective the best prompting formula so I'll give you one for the text and I'll give you one for the video but in the next lecture you will work on

58:17 it more in depth so I'll give you the basics and next lecture you will have a more deep understanding on how it works um so prompting formula made by shaunas digitas so uh what's important here is basically in the top what is less important at the bottom so the most important thing thing is obviously a task um and um again we have a context we have example we have Persona we have format we have tone and then uh details and restrictions are if needed so let's uh go a bit more in depth into each one so start with generate um give

59:05 a summary write a post and so on so this is a task it can be one task or a sequence of tasks so I'll give you a quick example um so let's say you want um a sports plan um so I want a sports plan so let's say I want a sports plan give me a sports plan for Monday to Sunday um so I'll give another another step so put it in a table Monday to Sunday and I want to have like a early morning exercise and then you know winding down uh stretching routine for the evening um and you know put that in the table um so that's you know basically um a task so it it's it's going to generate you a sports plan it's

59:51 going to do Monday to Sunday and it's going to do you morning and evening workouts uh which is you know uh Mobility probably in the morning and then stretching in the evening um another important thing is context so if we take the same example of you know me trying to get a better uh Sports plan so if I tell that generative AI tool that you know I have um I don't know a sprained ankle I have uh some you know hip mobility issues maybe you know I have some knee problems so if I put all of this context into this you know tool obviously all of these exercises that are going to be given to me are going to be adapted to me so context is very

1:00:38 important to get best results um also example so let's say you have a some sort of example uh of someone's you know exercise routine but you want to keep the format but change it a little bit for yourself then you can just simply attach it and by giving that example to your large language model to your chat gbt or whichever one then you obviously have to um um you know give that example and get better U results so Persona right uh like you are or write in style of so it really helps if you are you know aiming to write something you know like a speech or like let's say you're writing something um

1:01:24 more like art based or you just need like a something to guide you where you want to put your own ideas but just like a like a backbone of your whatever the thing that you want to write so you can obviously um invite your tool to act like as some sort of persona so if you want your text to be criticized by someone who is like a text critic let's say so you can ask chat GPT for example you know act as a Critic and you know criticize this work in some sort of manner so it really helps and the more Persona is known the better result you get because you know that person is online more and it's seen more so the you know the data is trained

1:02:11 on that person a little bit more so you it understands what you want more and another part of it is is um um a format so um it kind of um depends on your uh subscription but you can get um like let's say a HTML code a PDF a picture an audio uh or other format so uh you know you have to make sure that you ask for a format that you want so it can send you a PDF it can send you a jpack file it can send you a PNG file uh but you just need to ask for it it can send you HTML code or any other programming code that you probably know now already um also um lastly um a casual

1:02:57 tone a witty tone let's say more pragmatic and so on so um you can get you know your text uh more pragmatic or more informal so it depends on what you want so let's say again you need to do a speech wherever at your work or like you have some sort of lecture um so you can basically ask chbt or other tool to write you like a Qui quick resume about you which is a bit more informal or more formal depending on the format where you want to talk about um so um Indra uh is asking a good question in in the chat so if I write the prompt according to the recommendations that you give wouldn't

1:03:42 it be quite complicated for AI to understand that's a very good question um but this is where like generative AI tools are different so it's very important to understand that uh when we talk about Googling information so we have to know the keywords so when we are Googling we are writing only one word two words so basically you know Google algorithms work in that way but when we are trying to get best results from generative AI tools the more information the better that means you know that tool understands um you know that sequence that you are coming up with and is giving you the best answer so what I am giving to you here is just basically a formula like a like a structure uh of the sentence so and with

1:04:28 this structure you can have not bulletproof uh prompt but as best of the prompt as possible so I can quickly show you an example so we can understand a little bit better but again you will work on this more in depth uh with in in in the next lecture so again uh so this is a quick example uh of my perfect prompt so I am a 34y old male with a knee trauma that is context and gen generate a triathlon training plan that is a task for a month and put it in a table that's a task two plus format also give me some tips like you are a pro triathlete that's a Persona I wanted to act like a pro triathlete

1:05:13 write it in a simple terms and make it short so that's that's a tone that I am asking for chat GPT so when I ask this it kind of seems like a prolonged um um prompt but it actually works really well um so um let's um see uh some other chat GPT tips and tricks um but before going there let's let's actually see how Chad gbt will react to U this prom that I have so let's see uh if I can get some good triathlon uh training plan with this and um if you're up for some practice uh so I have a quick practice session to you which is one minute long

1:05:59 but uh um please try and write a prompt using this formula to this chat and see if you can uh you know learn from uh from your peers see how different people prompt see what you know different ideas of prompting can be um so um let's uh let's try it as well before I try and uh put this in in my chat gbd so try and write a prompt into chat or if you're watching online just for yourself in your chat gbt in your generative AI tool just try it out so basically you can understand it a little bit uh better so um I will try to uh uh copy this uh I will try to copy this uh

1:06:47 prompt and we'll we'll see uh what we are going to uh come up with okay sharing the chart GPT um so let's uh let's delete uh all of the formula parts so it wouldn't be uh in the way I hope you're uh trying to write your own prompt uh whilst I do this as well so let's see [Music] uh m so we already see some um really good stuff coming

1:07:36 out so not only we having this four-week triffon training plan uh also it's going to be partly for my uh knee trauma recovery but also I will have some uh Pro tips for knee friendly uh Triathlon training so one of the tips obviously is focus on the swimming more which really makes sense if you're trying to avoid more knee injury uh some low impact cycling uh some running with caution um only jog if your knee is painfree so don't jog if you have any problems prioritize recovery so stretch foam roll uh um KNE specific strengthening exercises listen

1:08:23 to your body and here we have you know a perfect plan basically that we can repeat for four weeks and hopefully we are good to go um again some a lot of endurance stuff um some drills for swimming and uh quite a bit more rest than what would you usually see for Triathlon training because you know Chad BT kind of understood that you are asking it to help you save your knees a little bit more so um it's uh you know quite U um quite different uh from what you what it would be simply without the context by the way so from my perspective again this is like a just a quick showdown of you know how this uh prom formula works but again you will

1:09:10 work on it more in depth uh um here so uh um I see that uh quite a few uh uh quite a few uh prompts here um I want to copy Natalia's prompt here just to check um what answer can we get for everyone I hope Natalia doesn't mind um but it's it's again it's it's quite important for you to you know understand how other people think how they prompt you know what can you learn from them as well um so it's it's it's it's pretty cool um that you can learn from each other as well so please do um so Natalia sends me a smiley face so

1:09:57 she's not angry I can tell so thank you Natalia okay so let's go back to um um learning uh more um about AI okay um so what I'm going to show you now is just um a couple of quick uh chat gbt tips and tricks uh so obviously women go created this uh uh big questionnaire throughout you know each and one of you and asked which you know uh tools are U mostly used and Chad gbt was you know way way uh above everybody

1:10:43 else all the other tools so I want to give you a quick tips on using chat gbt um and I'll try and show you again skipping from um my slides to chat gbt skipping to slides again to chat gbt so bear with me but I want to show you some quick Basics that might really really help you um to work on this um so um uh first of all uh we have some keyboard shortcuts that you might have not known um but it's quite uh you know simple uh it's quite simple thing um but it's uh it's very useful actually um so if we if we go to um our chat gbt uh if we go to our new clean window uh and if we go to um um this question mark at the bottom

1:11:29 bottom right corner uh so it's here uh if we click on that we have this um really sorry I did a Lithuanian uh uh Workshop uh today with teachers here so that's why it's in leanian it's usually in English but this is where you get you know um like a quick quick controls and if you click on that you you can see like all of these uh things that you can uh actually um actually do so I'll try and uh change it to English so you can understand it a little bit more so keyboy shortcuts you can open a new chat you can focus checked input you can copy uh last code block uh if you are programming so it's really useful and it's you know quite um

1:12:15 um um more comfortable to use it in that way so I can you know show you a quick example so let's say I want to translate something into let's say uh leanian translate a banana okay banana is bananas in leanian so it's a bit too easy but uh translate an apple into leanian so this is a quick leanian lesson to you guys um so uh at translation of an apple into leanian is obis so now you know obis in in in leanian is Apple um also uh translate uh uh also Translate um like a pair so a pair is uh leanian koua so that's how

1:13:02 you uh go up here but uh now if I um if I write uh let's say uh pineapple so it will automatically uh translate it because we are in the same window so because now we started talking to J GPT in the same window so now it kind of translates everything into leanian um but if we click on shift commo we are in an open new window and then we can you know try a different chat and then it will not be about translating fruits in tanian it will be a new chat so just keep in mind that when you open a new chat and you can do it with you know these um uh keyboard shortcuts that you are starting a new conversation but if we are keeping with in in the old conversation you are

1:13:48 you're having this old things that you already talked about as as a kind of a trigger to Chad gbt so if it's it's a bit more difficult to CH gbta to understand what else you want from it if it's especially if it's a new conversation that has nothing to do with translating you know fruits to leanian so just U keep that in mind so um another uh quick tip is basically that you can uh sort of sort your uh chats so let's say if I don't want this to be seen But I want to have it in in you know in my folder somewhere I can archive it so that means it's not seen here anymore but it's in my archives so if I go uh to my um settings I can you

1:14:34 know manage my archived child um um archived U chats and I can you know U restore it if I want so you can obviously rename them it's you know quite simple but if it's very important to you you can keep it here let's say translated names in leanian you can obviously do that um so it's you know it's quite simple things but it's just you know to understand that you know you can actually sort these things out and it will be a bit more um easy for you to understand you know what's what's there in your chat GPT uh you can obviously share and drag and drop to your notes so one of the uh things that I use quite often is dragging and dropping so let's say I want this in my notes I just simply drag it and drop it to my notes

1:15:20 then I have it and then when I click it I have it in my notes if I need to and of course I can share it so let's say I want to uh share my uh translated names um in thean or my Triathlon plan let's say so I can you know simply simply share that by creating a link and you know I can take that link and uh paste it into uh our chat and now you all have my uh Triathlon uh training chat so it's quite U again Simple Thing uh to u to do um you can also edit your um each of your uh prompts so if you go to uh The Prompt that you uh recently created um

1:16:08 you can uh just click edit and you will find this button here and you can let's say uh all over Africa not Europe and if I send it now it will you know recreate it uh you know the whole prompt and I can you know change my um whole picture so instead of you know me doing a lecture for European audience I I'll now be doing it to the African audience and I don't know if it's different or not but uh here we go it's a little bit different amazing women Tech so um yeah again it's a it's it's a useful quick uh tip um so um also

1:16:55 um you can use audio function uh and it's really um underused from my perspective uh so audio function obviously on your phones um but you know you can use it as a white piece of paper and why am I saying so um so um I'll give you a quick example of how I used my voice chat GPT on my phone so um I was learning to play blindfold chess which is basically playing chess not look at the board and um so I needed an assistant uh which would tell me coordinates in the board and I will tell that assistant whether that space in um you know this on this board is Black or White so I was

1:17:43 basically playing that with chbt and then said I like let's go to a Next Level so you will give me a position of a knight and I will tell you on which squares that Knight can jump and I was basically you know creating this little game of me learning blindfold chess um and I kind of created from scratch um and you can create it for anything else so you want the tutor to help you to learn more programming to be to become a better web designer um you want to you know ask you codes of colors uh you want to uh do like a quick test of something uh something to do with you know remembering what we just learned uh um if you do homework that is prescribed to you by the way um so so many uh you know

1:18:28 different things that you can think of you can actually do it via uh voice command so I really encourage you if you have it um use it it's it's really fun and you can learn from it quite a lot um so um another thing that I wanted to show you is um customizing your chat GPT so um I will quickly uh again show you um this part um so customizing chat GPT um it's quite uh quite important um because you can get again uh better uh better results uh from it that way um so um I'll quickly show you what I mean by that so again if you go to

1:19:14 chbt uh you go to your profile which is top right corner um and you go to customize chat GPT then what you get is basically some information about you and I didn't lie I'm 34y old male excited about technology and living in vill lethania um but also what I have here and this is much more important so I want that CH gbt to talk to me in in some way that I want to so I don't want that chbt to mention that you know that it's a large language model or it's AI I don't want that uh you know chbt um to respond um if it's unsure about answers uh I don't want that chbd to lie to me if it doesn't know the answer um I don't want that chbd to

1:20:00 apologize because um it is a large language model so it's all of these you know little things that would be a distraction to me I don't want chpt also to tell me that you know but please talk to a doctor or like someone a professional because I am an AI agent so you know it's it's a Time waster to me I want chbt to be specific and give me answers that I ask him to so uh this is is actually shared through many different uh technology enthusiasts um a same way to respond uh to to chbt is basically copied and paste so if you are you know um if you want to change it to what I have um I can quickly share it to uh chat and you can use the same one if you want or you can change it in

1:20:46 whichever way you want as well so again be mindful that you can customize the way chat GPT talks to you and it's quite um an important thing from my perspective at least okay so what we learned now is just a quick uh uh prompting formula for text um I want to give you again uh a quick understanding on how to prompt for the visuals but uh uh more you'll learn in the next session um so when thinking about prompting uh with visuals uh so we have to have in mind subject action environment details and attributes style and mood perspective and composition so these are the main things that we have

1:21:31 to keep in mind when we are prompting for a visual um so how it works so clearly State what the main subject of of an image is um describe what the main subject is doing or the main focus of the image uh when it's environment provide the details about where the image takes place so describe the background atmosphere and ambience and the details include any specific features of the subject or setting mentioned textures colors sizes and other descriptive attributes and style and mood so specify the artistic style of medium so you want it realistic you want it abstract you want it water color you want it to be I don't know anything that you can think of so you can play with these different styles and convey

1:22:18 the emotion or tone of the image um indicate the Viewpoint or angle so that's a composition part describe the layout and arrangement of elements in the picture so if we think about a perfect prompt for a picture so here it is subject plus action environment details and attributes style and mood perspective and composition so again this is from my perspective a perfect prompt so let's say it's a small playful cat which is a subject um batting a colorful toy which is a mouse that's an action on a wooden living room floor which is environment and sunlight streams through a window casting warm light and soft Shadows across the room which is details and attributes the cat's fur is a mix of

1:23:05 gray and white slightly fluffed with excitement while the toy mouse is bright

…