Up next in 10
A bit of AI - Episode 8
3K views
Nov 16, 2023
What is it like to work in the field of AI? How do you get started with AI? And what is going on this week? Find out in this 30-minute show hosted by cloud advocates Henk and Amy. In this show, we will enter a conversation with our guest who works with AI daily and challenge you to expand your skillset with a recommended weekly MS Learn module. Featuring Guest : Saba Samiei (https://twitter.com/SabaBigTechFan) C# Live - Dev Streaming Destination: https://www.c-sharpcorner.com/live C# Corner - Community of Software and Data Developers https://www.c-sharpcorner.com #abitofai #talkshow #podcast #csharpcornerlive #csharpcorner
View Video Transcript
0:30
Thank you
0:59
Thank you
1:29
Thank you
1:59
Thank you
2:29
Thank you
2:59
Thank you
3:29
Thank you
3:59
Thank you
4:29
Welcome to the A Bit of AI Show with your hosts, Hank and Amy
4:49
Hello everyone and welcome to the A Bit of AI Show. Can you believe we're already at the eight episodes
4:57
We've been doing this for two months already, Amy. Oh my goodness. No way. We were chatting with our guests before and I was like, how is this number eight? Also, we've been doing some planning on future episodes and guests as well. So yeah, it's been an exciting couple of weeks
5:15
Exactly. So welcome everyone to the A Bit of AI show. This show is all about the people in AI
5:24
So what did they do? What did they do? And how did they came into the field of AI
5:29
And we are collecting all these different stories from all these different type of jobs within the AI space
5:37
So today we have Saba. Let's just call her into the show and start our session today
5:47
sure before we chat with Sarah though let's just briefly say don't forget that anything we speak
5:55
about here on the stream you can chat with us after as well as if we're in person is the idea
6:02
and it's offline it's in a in a meeting environment where we have a code of conduct and it's called
6:09
our a bit of AI cafe so a nice place for you to chat with me and Henk if you want to but I'm
6:17
pretty sure you'll want to speak to our speaker instead, especially this week. So go to a bit of
6:23
a aka.ms, sorry, slash a bit of AI dash cafe. Perfect. So hi, Saba, welcome to the show
6:39
How are you today? Good. Good. Thank you. Kia ora, Kia ora, Amy, Kia ora, Hank from New Zealand
6:47
I'm good. How are you guys? Not too bad. Not too bad
6:54
It's getting summer here. Yeah, very good. So I'm happy. I know I'm jealous
7:03
And thank you so much for joining us. Obviously later into your evening as well
7:09
I'm sure you already had a busy week. Can you believe it? Only at Thursday
7:13
but we're nearly there nearly at the weekend for a bit of a bit of respite hopefully but it's an
7:19
absolute pleasure to have you so Sabah should we get started and one of the the first things that
7:26
we always do is kind of those that basic question that says tell us who you are and what you do in
7:32
the space of AI. Awesome cool so my name is Sabah Sami I'm originally from Iran and I live in New
7:42
Zealand. I am an AI and AI ethics researcher and enthusiast and I have founded two AI companies
7:52
one last year called Confident AI and one this year called Maxo
7:59
Very nice, very nice and we will dig into so much of what Sava just said there so bear with us
8:06
because there's lots of great stories for her to tell. So I guess one of the first questions that we always ask
8:14
that gives people that level set of what is it actually like to work in AI So all jobs in the industry look different daily but if you could describe an average day in your role
8:26
what does that look like? So I think anyone who has this founder role
8:37
and is kind of in the startup world would know that the days are actually quite different
8:42
And, you know, it starts in a different way and it ends in a different way
8:48
But I think if I were to summarize my day-to-day activities, it includes a lot of research, a lot of thinking and strategizing and talking, actually, to a lot of people
9:01
Sometimes at the end of the day, I come home and I'm like, my jaw hurts. but yeah
9:07
that's what basically what it looks like keeping up with the news
9:11
what's going on around the world that could impact us or that we need to take into consideration
9:18
when we are strategizing yeah talking to people, developing stuff coming up with ideas, validating them
9:26
all that kind of stuff cool So my question is, how did you get into AI
9:40
Okay. So I watched a movie called Transcendence when it was out
9:47
I think it was back in 2014, 15. And I remember vividly, quite vividly actually
9:56
that I went to the cinema with my ex and we watched this movie
10:00
and I was just amazed by this thing that happened, this human that was connected to the internet
10:06
and then all of a sudden became something completely different. And I think the trigger point for me was that just like the character
10:15
the wife, I was also like, oh, my God, he's come back to life. And then all of a sudden it changed and I was like, no
10:21
that person needs to go because that's now bad. And then I started, as soon as I came out of the movies
10:26
I was like, I'm going back to university. I want to continue my master's and finish my master's in AI
10:34
And my ex was like, yeah, right. And I proved him wrong
10:38
So, yeah, that's kind of like the actual trigger point. But I think the other one was the fact that I am a big fan
10:45
of Stephen Hawking's books. And I came across his warnings about artificial intelligence
10:52
And then kind of that made my love for AI a little bit more focused
10:57
and and i got into the ethics side of it which we can dig into that a bit more
11:03
wonderful i was gonna say no it's it's always interesting to see where um everyone meets ai
11:12
um and ai being such a big word for many many many things um so yeah it's interesting to kind of
11:19
share that some people come from the more technical side some people come from creative side
11:23
it's so interesting you come from the story side of it the the actual you know um the outcomes of
11:30
some of the things that we that we do in ai and so i'd love to ask you a bit more about that masters
11:36
because your master sounds like it was the the point in which you started to really dig into some
11:42
of this research so can you tell us a little bit more about what you studied yeah sure so um my
11:49
master's is actually in AI ethics. So I did a thesis-based master's, which meant I had no classes. I only had a research
11:57
for the two years that I did it. And it was about the why behind the danger
12:05
of AI. And the reason why I stumbled across that is initially I went to university and I went with an idea
12:13
of I want to create a device that when robots take over that it would kill all the robots but it wouldn't harm any humans
12:23
and and I went to that to the university being very very excited about my idea
12:28
and I was like bring me the Nobel Peace Prize and that didn't happen so the
12:35
the professor who ended up being my supervisor later was basically challenged me on well why do you think the robots are going to take over and
12:44
And why do you think these warnings exist? And so that kind of set me on the path to then understand why is it
12:52
and why are so many people kind of telling us that AI is dangerous
12:57
but at the same time investing heavily into it as well. So that was kind of the main topic of my master's
13:06
and it's called On the Danger of AI, if anyone is interested to go look it up
13:12
Wonderful. and um you in that research you mentioned um a promise that you would make at the end of it
13:22
can you tell us a bit more about what that was yeah um so in the uh basically the introduction
13:31
part just before i get into the actual kind of master research um i write about the fact that
13:37
how moved I was by doing this research, not just because of kind of seeing the possibilities
13:45
but the mistakes that we've made in the past and how we didn't learn from them and still kind of
13:51
moved forward with some of the other developments that we have. And so the promise that I made to
13:57
my readers was for them to be a person that has read the work of someone who went on and made a
14:06
change in the world and because that was written and it was published now I'm holding myself to
14:12
account to actually deliver that and and that's the story behind how Maxo and comfort AI came
14:21
about as well that's my way of making that impact. Wonderful yeah I know that's it's it's always
14:28
great right when you see research take that that next step as well and become something that like
14:35
you said can really grow into something so could you mention comfort ai and i think comfort ai is
14:42
one place that we kind of wanted to start can you tell us a little bit more about what it what it is
14:46
sure so one of the things that i realized in my master research is that um this fear of artificial
14:57
intelligence or part of the fear or part of the danger as well uh is because people don't really
15:04
know what it is. They don't know the possibilities. They don't know what it can and cannot do
15:09
And they don't know by feeding these AI engines that we use on a day-to-day basis, then the impact
15:15
that they're making on the future of not just themselves, but the future generations as well
15:20
So the first thing I started to do was to start educating people around me, as many people as I
15:26
reach and that's what comfort AI kind of came from as well I delivered a workshop to high school
15:34
students here in New Zealand and they loved it and then I realized how much how many questions
15:40
they have you know that next generations there were generations then how many questions they
15:44
actually have about their future and so I started designing these workshops for high school students
15:51
as well as people who are already in the workforce, to teach them about artificial intelligence
15:58
And the promise or the tagline for Comfort AI, which is my personal mission as well
16:03
is to make artificial intelligence the comfort zone of as many people as I can reach
16:08
And that just means helping them actually understand what is AI, what are these different buzzwords that we use in AI
16:16
and then what does it mean practically for them in their day-to-day life
16:21
Yeah, it's, interestingly, we've had people on the show who have mentioned, we'll ask
16:27
you a question actually shortly that might test you a little bit. But we, we chatted somewhere and they said the expectation is often something that needs
16:37
that managed within AI So we see a lot in all sorts of different places including you know media And it like is that what it is right now
16:51
How can we use it safely? How are we thinking about what we can do
16:58
as people who are creating technology to build responsible AI? And so you mentioned these workshops
17:06
and you mentioned the age range in which you share them with, that's really exciting
17:10
that there were sort of high school students, so teenagers and young people
17:16
How have you found people kind of coming on that journey? Because AI, to a lot of people, seems very difficult area
17:24
of technology, right? There's a lot of complexity in that. How are people finding that learning curve
17:31
I think once you tap into how AI or the role that AI plays in people's day-to-day lives and the real examples that they're dealing with, then it makes it a bit easier for them to understand
17:47
I mean, obviously not everyone's going to be a developer. And I think that's one of the mistakes that the AI community sometimes makes is that we
17:54
are so kind of tied up in our own world that when we talk to others, we might come across
18:00
as like aliens. So I think for me, it's kind of been like, do this the other way around
18:07
That said, for that workshop that I did and a couple others that I did after that, I went
18:13
to that workshop thinking, okay, well, I'm going to tell them about the definition of artificial intelligence and the definition of machine learning, blah, blah, blah
18:21
And then I went to that workshop and I said, well, you guys tell me what is it that you
18:25
want to hear? What is it that you want to learn? What's the one thing that you want to take away from this workshop
18:31
And the first one asked me a question, you know, what is different languages that are
18:36
used in machine learning, for example. The second one asked me, can we replicate our consciousness into the cloud so we can
18:42
live forever? and as soon as i heard that i was like okay this generation knows a lot more than we think they do
18:51
so um and i was so glad that i did my masters and i did some research on that topic as well
18:58
consciousness and so i could actually guide them the outcome for me was very interesting even though
19:04
that is not what i intended um at the end of the workshop i did ask every attendee to come and you
19:11
Tell us what is it that they have learned. And almost every single one of them, and this is not an attack
19:17
on any particular company, this is purely for having educated people about how their data is being used, almost every single one of them
19:26
decided to close their Facebook account after that workshop. I don't know if they did it or not, but that was one of the things
19:33
that happened, and I found that quite interesting. Some of them decided to go back and challenge things
19:40
that were given to them. So that was another moment that I was quite happy with
19:45
And by challenge, I mean, you know, not just accept anything that's given to you
19:50
Just go and do your own research and really understand the reasoning behind some of the things that are said
19:57
So, yeah, I think that was the story behind the workshops. And I founded that in February 2020, which we all know how 2020 was
20:07
so then I converted that into um so I couldn't do as many workshops as I wanted but I converted
20:12
that into a YouTube channel uh and and a an Instagram Twitter and LinkedIn channel as well
20:19
so I post videos on it and an educational material but I'm going to get back into doing
20:23
the workshops now that we're kind of out of the lockdown wonderful yeah no a difficult time um
20:31
Definitely. I know. Well, we're here obviously online doing a lot more online events as well
20:38
So it's exciting to see that you're able to transition and still be able to offer your advice in the area
20:46
So, Henk, you have our always tricky question, but really telling question that you ask our guests
20:56
Yeah, I think we already touched a little bit about it on it. But what is the most annoying thing about your role in AI
21:05
Yeah, I think we did touch a bit on that. It is expectation management
21:11
I feel that people underestimate the long-term potential or the danger in a particular AI product
21:24
and overestimate what can be achieved in the short term. So when I have conversations with different people
21:33
about artificial intelligence, it's normally either people who are really scared of it or people who are very disappointed by it
21:45
So those who are really scared of it, and by scared of it, I mean those who think, you know, it's going to become the Terminator
21:50
and kill us all, they're kind of like, they're basically me before I did my masters
21:58
And I don't blame them for thinking that either. I just think that now that I have done my research
22:05
that that danger is actually beyond that. But the other group of people who are either
22:09
who end up being disappointed about artificial intelligence are those who are normally like investors
22:14
or see suit of a particular company that were promised that we can do such and such with AI
22:21
and we can do it in such a short time. And then someone did something for them and it didn't work
22:26
and then you only get one shot at the first impression type thing
22:30
then changing people's mind after they have already experienced something is quite difficult
22:35
And then as a founder, I need to talk to different people
22:40
about what we can and cannot do within the given timeframe. So, yeah, I would definitely say expectation management
22:46
is one of the challenging parts. I don't want to say annoying
22:51
Yeah, no, challenging is fair. Challenging is a fair word. and we've heard that from a couple of our guests
22:58
and so interesting that even different roles within the same space are managing expectations
23:05
So thank you so much for sharing your story and Comfort AI and the education that you're doing
23:12
So this is really exciting. We are going to move on to something called
23:16
our quickfire round. So Sabah, we have briefly touched on what's going to happen here
23:22
but we have not told you the questions in which we will ask. And so the ask of you is to answer as quickly as possible, as briefly as
23:32
possible, and with the first thing that comes to mind. And we have about six questions for you
23:39
So I'll start. So nice, easy one to get us started. What was your first programming language? C
23:47
Oh, okay We've had some Low-level device hardware Program languages That's been a bit of a pattern
23:57
Now thinking about it Yeah And what was the programming language You used last Python
24:10
Oh, no Oh, MATLAB actually, I did a machine learning course. Oh, nice
24:20
Yeah, interesting, good stuff. That's a, yeah, still a new one. Good stuff, thank you for sharing
24:27
So what was the last thing you learned in AI and if you can keep it to no more than a sentence
24:32
Oh, the last thing I learned about AI
24:44
It's data. It's not AI. It's okay. I'll go quick. This is the first thing that comes into my mind, that 92% of children under the age of
24:57
two in the US have a digital footprint It not AI related Wow Interesting But data fuels doesn it So that fine We will accept We take that answer That an easy one Cool
25:10
Favorite event on the AI calendar? Singularity University events always fascinate me
25:27
What area of AI is on your list to next scale up on
25:31
So what are you looking to learn next? Machine learning. Fun. Nice. Nice
25:39
And what training framework did you use last? Can also be none
25:49
Training framework, like where I did my training from? No, like PyTorch or TensorFlow or psychic learn
26:03
I can't. Oh, that's a hard one. This is, you should have given me so far
26:11
I can't remember. No, that's cool. That's cool because it shows us all the very different perspectives that happen in AI
26:20
So as much as we are very deeply technical people who work every single day with frameworks, then we have people focusing on data, we have people focusing on ethics and research
26:30
So that's it's a great answer. It helps us to understand the broad variety of different things happening in this space
26:38
But that is pretty much all we've got time for. One of the things I did very briefly want to ask you, actually, was just about your learning ethos
26:48
So everyone that we have on the show tends to chat a lot about how to keep up with the ever changing environment
26:57
Do you have a way that you learn? Can you share with us
27:01
Is it is it reading? Is it doing? Is it, you know, et cetera? Yeah, it's it's very hard to do that, actually
27:09
I know. So I do a few things. I read books. I read as much as I can, whether that's like actually reading books, which I love, or just listening to audio books
27:21
I do a lot of online courses. So I think that the reason why I asked the online module, I use Coursera a lot and also Microsoft Learn
27:33
And I read a lot of articles. I think I find those very useful because they're short, they're kind of very summarized and, you know, I don't always say I don't just stick to the headline
27:47
But then sometimes when you're like reading something very quickly, it does come with like bullet points as well
27:53
So if you're in a real hurry or you've got a lot of things to get through, then those short articles always really help
27:59
But I do recommend that anyone that reads an article read it from a few different sources, especially if it's a news article
28:07
Yeah, no, it's always good to get multiple perspectives. Same within the air, I suppose, right
28:13
Building teams of multiple perspectives, building ideas and asking people their feedback is always important
28:21
So thank you so much, Sabah. It's been a flash. Can you believe that much time has already gone
28:27
But it's been so interesting to hear about kind of what you're working on, how you've got into AI, your specific area of AI as well
28:37
And we really appreciate your time. So thanks so much. Thank you so much for having me
28:43
This has been great. Wonderful. Well, we will see Sabah in the cafe
28:50
Exactly. Yes. I'll be there. I see some questions in the chat and these are typically the questions you can ask in the cafe
29:00
So please join the cafe in about five minutes. Thank you for joining, Sarah
29:07
So that's a bit of AI-cafe. Bye, Sarah. So, and we are already at the last part of our show
29:20
So we have two things left for you. We have the event of the week or month, or maybe even year for you
29:28
And we have our learn modules. So I think the big news is that registration for Microsoft Build opened this week
29:38
And that you can now register. And there will be a lot of announcements
29:43
So maybe there will be some cool announcements around AI. Who knows
29:48
so please go to mybuild.microsoft.com and register for all the sessions and it's free
29:57
so learn module of the week time wonderful yes this one has been an interesting one
30:07
obviously we've gone down the route of ethics and responsible AI and so two of the
30:15
modules that I wanted to pick up on are things that can support you as you're looking at building AI products and understanding things
30:24
like what is our model doing, as well as trying to detect and mitigate unfairness. And so the two
30:30
modules that I've got are explain machine learning models of Azure machine learning, as well as
30:35
detect and mitigate unfairness in models with Azure machine learning. And so if you want to
30:41
We've set up a challenge for you. It's at aka.ms slash a bit of AI dash learn
30:48
so that should be open right now. Lots of people tend to go in there
30:52
and complete it from week to week, so thank you for participating as always
30:57
These are great modules that can really show you different parts of the Azure
31:02
Machine Learning space as well as, you know, some great learning to
31:07
kickstart you in this area so that in this area so that you can learn more about these frameworks because that's some open source
31:14
which is great um so that was kind of our piece on learn we really appreciate you taking part uh
31:21
always you know keep in touch give us feedback that's what we like to hear um and so that we can
31:27
make sure that we're building great content but really um that is all from us hank already today
31:35
So thank you. Thank you for joining us. And we hope you've enjoyed the show. We're here every Thursday at 10 a.m. Central European time and all different times across the world
31:48
If you go to our website, a bit of AI dot show, you can see lots of different time zones in which it's in. So that's incredibly useful
31:55
and yeah please join us in the a bit of ai cafe after the show so that's aka.ms
32:03
slash a bit of ai dash cafe if you want to re-watch this episode or grab any links you
32:09
can go to our website a bit of ai.show um and hank any anything you wanted to add anything
32:16
special this week no but in like 24 hours it will also be available as a podcast
32:24
So you can have us while you're walking or in the car
32:31
That's wonderful. We don't need to be directly looking at screens. There's always a good way to just have 30 minutes
32:40
listen to a conversation. It's the way we've designed it to be an engaging conversation
32:45
So great shout, Hank. That's a very, very good one. So thanks everyone for watching
32:51
Thanks everyone for watching. This has been a bit of AI show with Henk and Amy
32:57
We will see you next week. See you next week
#Computer Science
#Intelligent Personal Assistants
#Jobs & Education
#Machine Learning & Artificial Intelligence