00:00:00
[Music]
00:00:17
okay so for today I'm going just to give
00:00:20
a little introduction uh about the topic
00:00:24
and then since um I was told to discuss
00:00:27
about NLP I would also like to discuss a
00:00:30
bit about the difference between NLP or
00:00:33
natural language processing and the Gen
00:00:36
AI or generative AI ethical use of AI or
00:00:40
NLP efficient use again allow me to use
00:00:43
it uh in two areas uh education and in
00:00:46
the workplace no specific industry and
00:00:48
when we say industry I won't talk about
00:00:51
its application in the hospital or in
00:00:54
banking or in um let's say oil and gas
00:00:58
uh Etc so I'll choose education and work
00:01:01
because I think uh some of the a good
00:01:03
percentage of the attendees are students
00:01:05
and the others are newly graduates as
00:01:07
well and a future of NLP in the industry
00:01:11
what is uh NLP NLP or the natural
00:01:14
language processing it's a field of AI
00:01:18
that focuses on the interaction between
00:01:21
the human and human language and uh the
00:01:24
computers so it involves processing and
00:01:28
analyzing large amount of natural
00:01:31
language such as data or such as text
00:01:34
and speech and that it extracts the
00:01:38
meaning and enable the communication
00:01:40
between the humans and the Machine I
00:01:43
will not be talking about very technical
00:01:45
details about NLP I believe and I've
00:01:48
mentioned this to Richard let the
00:01:51
technical one be mentioned by the next
00:01:53
uh be discussed by the next technical
00:01:55
speaker so how does the NLP work and in
00:01:59
n LP excuse
00:02:02
me uh this involves breaking down a text
00:02:05
into uh smaller components like words or
00:02:09
phrases or sentences so it understands
00:02:12
grammatically the the structure or there
00:02:15
is a certain syntax and act actual
00:02:18
meaning or the um the the semantics so
00:02:21
the purpose of this is um it helps AI
00:02:25
models accurately process the text so
00:02:28
NLP you this is the mo uh is a mod model
00:02:32
used to accurately process the text it
00:02:34
identifies keywords the relationship
00:02:37
between the meanings and it facilitates
00:02:40
a better comprehension of the human
00:02:42
language so even if you say let's say
00:02:45
the the correct um sentence is my name
00:02:50
is Irene Corpus even if you say name
00:02:53
Irene Corpus it understands something
00:02:55
like that okay so the in context of
00:02:58
learning this reper refers to the nlp's
00:03:01
capability to enhance its understanding
00:03:04
by learning from past data or
00:03:07
interaction or conversations it allows
00:03:10
AI to interpret the context of the the
00:03:13
sentences it adapt responses responses
00:03:16
also based on interactions and it
00:03:19
improves the accuracy and relevance over
00:03:22
time so the the NLP is a a broader field
00:03:27
it's a broader field of AI and it is
00:03:29
focused again on interpretation of human
00:03:33
language it can interprets it such that
00:03:36
the machine and the humans gain um
00:03:40
develop an understanding of what they
00:03:41
are talking about it it converts one
00:03:44
language to another even for that one it
00:03:47
can do there's a speech recognition it
00:03:50
converts broken words into text text to
00:03:53
speech as well converting uh written
00:03:55
texts into SPO uh spoken words um what
00:03:59
are are examples of this um uh NLP tasks
00:04:05
classification sentiment analysis so if
00:04:07
you're going into some website you will
00:04:10
see a small icon at the bottom right
00:04:12
like the like a robot that has a uh a
00:04:16
headset no this is the the chat bot so
00:04:20
sometimes if you are interacting with
00:04:23
the uh with the chat bot the way you
00:04:27
enter let's say your text it can analyze
00:04:30
whether you are happy you are sad you
00:04:32
are frustrated this is what it says it
00:04:35
means when you say sentiment uh analysis
00:04:38
it can do language translation as well
00:04:40
and summarization so the NLP can use
00:04:43
both a statistical method and deep
00:04:46
learning to process your language it
00:04:48
makes uh making it a fundamental uh for
00:04:51
a range of applications including
00:04:53
grammarly and the Google translate or
00:04:56
translate.google.com and other search
00:04:59
engines
00:05:00
okay so what are these examples of NLP
00:05:03
uh models you have the voice assistant
00:05:05
like Alexa if anyone of you own the
00:05:08
Alexa you know that when you say Alexa
00:05:10
play the um Spotify and then it will
00:05:13
play your Spotify there's also the Siri
00:05:16
or the Google Assistant sometimes even
00:05:19
if you just put your phone near you and
00:05:22
you speak there is a Quee wherein the
00:05:26
this uh Voice assistance will initiate
00:05:28
the language transl
00:05:30
tool you have a Microsoft translator and
00:05:32
you also have a Google translate it's
00:05:35
translate.google.com
00:05:36
no at the a very early time when the
00:05:39
Google translate was implemented or
00:05:42
launched it doesn't have it it creates a
00:05:46
poor translation sometimes as since I'm
00:05:49
I'm uh Comm I communicate with um uh
00:05:52
various offices globally and I see a
00:05:55
communication that is in different
00:05:57
languages I copy and paste it in the
00:05:59
Google translate but when it translates
00:06:01
it it's like you won't understand
00:06:03
anything but right now um it is very
00:06:07
high in terms of accuracy of translation
00:06:10
um you can read you can understand the
00:06:12
context of the discussion because it
00:06:15
creates a better uh translation at this
00:06:17
time uh especially here in the UAE I get
00:06:21
some Communications that are written in
00:06:22
Arabic not only in Arabic and then
00:06:25
English text but Arabic language itself
00:06:28
when I paste it in the Google Now I'm
00:06:29
able to uh to understand the content
00:06:33
another um NLP tool is the text
00:06:37
analytics I've um I'm not sure whether
00:06:41
you have heard of the IBM Watson um the
00:06:44
IBM Watson um there's also uh AWS
00:06:49
comprehend and the Microsoft assure
00:06:52
texts uh analytics and uh there is the
00:06:56
um sentiment anal analysis API or um
00:07:01
application program interface like the
00:07:04
monkey learn or alien chatbots I've
00:07:07
mentioned earlier not only in some but
00:07:10
there are a lot of applications right
00:07:12
now or web uh websites instead of having
00:07:16
a real human to interact with the
00:07:19
customers now they Implement an AI
00:07:22
chatbot this um allows the company to
00:07:27
have a simplified communication
00:07:30
sometimes we see it as a simplified
00:07:32
communication because there are already
00:07:34
predefined questions or the FAQs and the
00:07:38
chat bot will um reply according to the
00:07:41
predefined question even the question
00:07:44
sometimes you will see how can I help
00:07:46
you today and then there are already
00:07:47
options of the questions if you click on
00:07:49
that there is already a predefined uh
00:07:52
answer however if your answer is not
00:07:55
your question is not satisfied in the
00:07:57
end there hopefully
00:07:59
not all of them but some of them will
00:08:02
still say option give an option to talk
00:08:04
to a um to talk to a human and that is
00:08:07
the time where you will be engaging with
00:08:09
the
00:08:10
humans so these are the V virtual agents
00:08:13
and other NLP tools the grammar and
00:08:16
spell spell checker I use grammarly when
00:08:20
I am creating um formal emails or when
00:08:24
I'm documenting let's say policies part
00:08:27
of my work is policy cyber policy
00:08:29
development so before uh before I submit
00:08:33
it for review I make sure that it is
00:08:37
properly uh written uh grammarly correct
00:08:40
even the punctuation the comma the
00:08:42
semicolon these are very important when
00:08:45
you are creating a document that will be
00:08:47
published in public and be used uh by
00:08:50
the public and in fact even the grammar
00:08:53
Lee now can be integrated with your
00:08:56
msword or your uh word processor no uh
00:08:59
there is an option also even without the
00:09:03
grammarly in Ms word there is an option
00:09:05
to review the grammar uh or spell
00:09:08
checker so that's is for NLP but before
00:09:11
I move forward there I wanted to put
00:09:15
something that relates to GPT because
00:09:18
they are similar but there is still a
00:09:21
difference but there is a difference
00:09:23
between NLP or natural language
00:09:26
processing and the GPT now GPT let's
00:09:29
look at the Sprints to 1 million users
00:09:33
Netflix took 3.5 years to reach 1
00:09:37
million there's also the kickstarter 2.5
00:09:40
and Twitter two years before it reached
00:09:43
1 million
00:09:44
users the Facebook it took 10 months to
00:09:48
reach 1 million
00:09:51
users Spotify it reached five months to
00:09:56
reach 1 million inst gram 2.5 months
00:10:02
okay so it took five days only when chat
00:10:06
GPT was launch officially launch to
00:10:09
reach 1 million users Okay now what's
00:10:13
the difference between NLP and
00:10:15
generative AI chat GPT is just one of
00:10:19
the generative AIS NLP interaction
00:10:23
between human language and computers so
00:10:25
it involves processing analyzing large
00:10:28
amounts of natural language data such as
00:10:31
text and speech and extract the meaning
00:10:34
and enable communication between the
00:10:36
humans and the machines no the
00:10:38
generative AI on the other hand it the
00:10:41
focus is that it specifically creates a
00:10:44
new content including text images or
00:10:48
audio and code it
00:10:51
emphasizes um on producing new outputs
00:10:55
rather than just understanding or
00:10:57
analyze the existing input
00:10:59
so functionally it typically uses deep
00:11:03
learning like
00:11:05
Transformers the GPT so whoever
00:11:09
mentioned generative pre-trained
00:11:12
Transformer then that's what it is a
00:11:15
chat GPT is a a gen AI it generates
00:11:19
coherent and um
00:11:22
contextually relevant text there are
00:11:25
there is also um model uh another de
00:11:29
learning model like the diffusion model
00:11:32
uh it is used in generating images and
00:11:34
other visual uh media um for those of
00:11:39
you who are using the chat GPT you will
00:11:42
see on the upper left side that if you
00:11:45
click one icon or button there it will
00:11:48
show you other um gpts that are
00:11:52
integrated with the open
00:11:55
AI uh you can see D doll e for image and
00:11:59
there's also Juke um jukebox I think J
00:12:03
Jun box whatever for music so there's
00:12:06
the Gan or generative adversarial
00:12:09
Network this is used for image syn
00:12:12
synthesis video creation and other types
00:12:15
of creative content Let's uh talk about
00:12:19
what else the fake the fake technology
00:12:21
so how it works it uses the gun by the
00:12:24
way Gan for deep fake it creates highly
00:12:27
realistic uh uh synthetic images or
00:12:31
videos of people make it making it
00:12:33
appear as if someone is saying or doing
00:12:37
something that they
00:12:38
haven't again make yourself um aware or
00:12:43
educated read and listen to the news not
00:12:47
only in the country not only in the
00:12:49
Philippines in the region Asia region
00:12:52
and globally you may have seen a lot of
00:12:55
deep fake that has been uh done with uh
00:12:58
let's say President Obama former
00:13:00
president Obama um Bill Gates um a lot
00:13:05
of other celebrities the Gen the this
00:13:09
deep fake are use are using the gun or
00:13:12
generative adversarial Network in the
00:13:16
gun there is the
00:13:19
discriminator the generator and the
00:13:22
discriminator so the generator this is
00:13:25
the one that creates the fake images or
00:13:28
or videos of the person while the
00:13:31
discriminator this evaluates the
00:13:33
creation this creation to distinguish
00:13:36
them from the real vide so the two
00:13:38
models they comp they compete and
00:13:41
improve uh over time and unfortunately
00:13:45
it's resulting in increasingly realistic
00:13:48
dip fakes
00:13:50
um let's see the next
00:13:53
one So Co covering this very important
00:13:58
for us to know is that there are ethics
00:14:01
and principles in Ai and these are
00:14:03
ethical consideration in AI
00:14:06
integration
00:14:08
um there are many companies now who are
00:14:11
adopting and implementing the AI
00:14:14
integrating them in their uh in their
00:14:16
day-to-day work there is there was the
00:14:20
recent gitex it's a one week technology
00:14:24
event and a lot of applications AI
00:14:27
applications have been launch there
00:14:29
mostly by um either private or
00:14:32
government entities and some of these
00:14:34
are let's say procurement applications
00:14:36
really designed for them and uh HR
00:14:40
applications so this HR applications
00:14:43
they already embed the information from
00:14:46
the employees it
00:14:48
analyzes um the the necessary training
00:14:52
required for the employee especially
00:14:54
after the performance appraisal at the
00:14:56
end of the year it should come out with
00:14:58
the the training that would fill in the
00:15:01
Gap in skills and knowledge of the
00:15:03
individuals also that is based on the
00:15:06
objectives and let's say the job
00:15:09
description that's written for the
00:15:10
employees so some of the ethical
00:15:12
considerations when you do AI
00:15:15
integration are um human Centric
00:15:20
um although AI has the potential to
00:15:23
significantly enhance decision or human
00:15:27
decision uh AI becomes more integrated
00:15:30
into our daily lives it is crucial to
00:15:32
ensure that it is designed and
00:15:35
implemented but aligned with this key
00:15:38
principle um in h being in human a human
00:15:43
Centric AI or integrated with human
00:15:47
Centric principle in mind and let's give
00:15:49
an example an AI assistant designed to
00:15:52
Aid doctors by providing
00:15:55
suggestions they provide suggestions not
00:15:58
make
00:15:59
the decisions it still keeps the control
00:16:02
in human hands so it it ensures this
00:16:05
human Centric IT
00:16:06
addresses um that the AI complements the
00:16:10
human roles only enhancing the safety
00:16:13
and decision making rather than
00:16:15
replacing the human
00:16:17
judgment transparency and
00:16:20
explainability uh what can we give here
00:16:22
as an example a financial AI tool for
00:16:26
example it shows users what why a loan
00:16:29
was approved or
00:16:32
denied explaining the criteria and the
00:16:34
logic for for this uh decision so it
00:16:38
helps users understand an AI decision
00:16:41
increasing the trust and enabling also
00:16:44
informed decision making so if you
00:16:47
notice decision making still lies into
00:16:50
the hands of the
00:16:52
humans fairness and non-discrimination
00:16:55
this is what's very important for me an
00:16:58
AI for for example hiring tool I
00:17:00
mentioned about uh HR earlier so an H an
00:17:04
AI hiring tool um audit audited to
00:17:08
ensure it doesn't favor any gender so
00:17:11
this is an example it doesn't favor any
00:17:13
gender any race or background in
00:17:17
candidate
00:17:18
selection very familiar T that if you
00:17:21
submit your CV it has your picture to me
00:17:25
I don't need to put my picture because
00:17:27
at this time you should know that
00:17:30
whether it is AI or whether it is human
00:17:33
your looks or your gender or your your
00:17:37
ra race or your nationality even your
00:17:39
religion should not be the basis for you
00:17:42
to be hired in a position the same thing
00:17:45
for AI it should be fair and
00:17:50
non-discriminative it prevents biases it
00:17:52
should prevent biases that could lead to
00:17:54
unfair treatment or discriminatory
00:17:56
outcomes privacy uh and privacy
00:18:01
protection uh safeguarding privacy and
00:18:03
ensuring robust data governance
00:18:06
practices example uh a health care
00:18:09
chatbot uh it anom anonymizes user
00:18:13
information to prevent misuse of
00:18:15
personnel Health
00:18:17
Data uh for this one IT addresses the
00:18:20
protection of user data from
00:18:22
unauthorized access and ensure
00:18:24
compliance with data protection
00:18:26
regulation in EU we have the GDP in EU
00:18:30
there's the gdpr general data protection
00:18:32
regulation in the Philippines there's
00:18:34
also the data protection uh law or
00:18:36
policy um it was given or mentioned at
00:18:40
the opening of the uh of this
00:18:44
webinar accountability and oversight so
00:18:48
uh this me means clear accountability
00:18:50
mechanism for AI advocating for
00:18:53
responsible a uh entities to be
00:18:56
identifiable and held accountable so
00:18:59
example is an AI surveillance system
00:19:02
with the human review process to Monitor
00:19:05
and verify any flagged Behavior or
00:19:08
alerts IT addresses um the um or ensures
00:19:12
that responsibility for AIS action
00:19:15
allowing for correction of errors and
00:19:17
accountability in deployment this one if
00:19:20
I may just in um inject something
00:19:23
something that I read just a few just an
00:19:26
hour before the webinar started there's
00:19:28
a 14-year-old uh teenager in the US who
00:19:31
committed suicide and the lawyer or the
00:19:34
parents now are suing the
00:19:37
platform because um it turned out that
00:19:41
the
00:19:42
14-year-old started engaging or
00:19:44
interacting with a chatbot so the the
00:19:47
teenager is a boy and started
00:19:50
interacting with a chatbot who happens
00:19:53
to be a character of a female and then
00:19:57
cut it short he fell love with the
00:20:00
chatbot and then you can just Google
00:20:02
that please uh can AI be blamed for the
00:20:05
suicide of 14-year-old try to read the U
00:20:09
uh the the news it will show there that
00:20:11
the last chat that he had with the
00:20:14
chatbot is that I hope I can come to you
00:20:17
now and the chatbot said I'm waiting for
00:20:19
you if you can come now something like
00:20:22
that okay try to Google
00:20:24
it fresh
00:20:26
news so
00:20:29
AI in general so there are distinct
00:20:32
considerations reason why I talk about
00:20:34
nle first and then the GPT is to
00:20:37
determine the distinction between the
00:20:40
two AI um uh generative AI again is
00:20:47
there's a Content authenticity
00:20:49
generative AI um has a concern about
00:20:53
misinformation deep fake or fake
00:20:56
creation the in terms of usage
00:20:58
restriction there must be Safeguard to
00:21:01
prevent the use of gen AI in creating or
00:21:05
uh creating misleading or harmful
00:21:07
content that's why the uh the
00:21:09
responsibility or R AI responsible Ai
00:21:13
and not only the use but also in
00:21:15
development is very important on the
00:21:18
other hand the
00:21:20
NLP the consideration there is the
00:21:22
accuracy in interpretation NLP needs to
00:21:26
be accurate and the and context aware to
00:21:29
avoid misinterpretation especially in
00:21:32
critical applications like legal or
00:21:34
medical text
00:21:35
analysis I um I recently actually just
00:21:40
got out from the hospital and when I was
00:21:43
reading all of the results especially
00:21:46
the uh the blood test the ultrasound and
00:21:49
the other tests it mentions at the end
00:21:52
that the result is
00:21:55
generated through voice recognitions so
00:21:58
when when the doctor was doing let's say
00:22:01
an ultrasound or a city scan it has a
00:22:04
microphone it talks it speaks and
00:22:06
interprets what he or she was saying
00:22:10
during the process of the ultrasound or
00:22:13
the city scan and that gets translated
00:22:15
to a medical report but at the end it
00:22:18
will say that that was
00:22:21
generated by uh um a text a
00:22:24
transcription or a voice transcription
00:22:27
it is very important and then that the
00:22:29
doctor will have to read and evaluate
00:22:32
the content of that medical report now
00:22:35
let's look at secure AI integration in a
00:22:38
in uh education and
00:22:40
workspace when we say secure AI
00:22:43
integration and education in education
00:22:46
and the workspace it depends on how you
00:22:48
are implementing the AI so do you
00:22:51
implement it um coming from the platform
00:22:54
itself or you're going to implement it
00:22:56
in your own platform or in your own uh
00:23:00
environment AI in application in
00:23:04
education uh can um for you for the
00:23:07
students AI can assist in um providing
00:23:10
personalized learning experiences for
00:23:13
the uh for the faculty it's automatic
00:23:16
grading and also improves actually um
00:23:22
student engagement however AI must be
00:23:25
integrated securely to minimize the risk
00:23:28
so AI application in the workspace can
00:23:31
automate on the other hand so education
00:23:34
let's say in the workplace it can
00:23:36
automate routine tasks improve
00:23:39
productivity and even identify patterns
00:23:41
in data in cyber security we have those
00:23:46
of you who are working in it it
00:23:48
operations or security operations
00:23:50
there's what we call sock Security
00:23:52
operation Center most of the tools that
00:23:55
we are that are being used in there are
00:23:57
are already a AI enabled like the seims
00:24:01
uh security incidents and events
00:24:02
management it already correlates data
00:24:06
and gives you a meaning meaningful
00:24:07
Insight if there is anything happening
00:24:09
in the environment it will tell or
00:24:12
mention that there is something
00:24:13
happening um therefore it is only now
00:24:17
the the uh the security officer's
00:24:19
decision to look and analyze that data
00:24:22
but definitely the bigger part of data
00:24:24
analysis and the patterns is already
00:24:27
done by the by the AI so this improves
00:24:30
productivity and more um more efficient
00:24:34
in addressing uh
00:24:36
situation AI also in workplace has to
00:24:40
have a corresponding AI use policy if
00:24:44
any one of you
00:24:45
here company are implementing AI make
00:24:49
sure that you have a governance and
00:24:51
policy in place in uh last year in
00:24:55
April Samsung Electronics there was an
00:24:57
employe
00:24:59
who
00:25:00
inadvertently leaked sensitive company
00:25:02
information by using chat GPT and we
00:25:05
know chat GPT uh it still continuously
00:25:09
improves no if there's anything that is
00:25:13
let's say unus unusual and reported by
00:25:15
the end users they listen to that and uh
00:25:18
one of the things I heard as well in the
00:25:20
chat GPT is someone complained that why
00:25:22
is somebody else's discussion or
00:25:25
communication coming out in my GPT
00:25:28
you're checking in or logging in when
00:25:30
you're using chat jpt so you are
00:25:32
expecting that most Communications are
00:25:35
coming from your previous entries as
00:25:37
well but at that uh at a certain point
00:25:40
somebody complained that or raised a
00:25:42
concern that somebody else's discussion
00:25:45
is coming into his GPT so that case in
00:25:49
the Samsung the there was an engineer
00:25:52
who input who input a confidential data
00:25:56
including source code internal meeting
00:25:58
notes into the AI tool um it led to the
00:26:03
information being stored in the external
00:26:06
server so in response the Samsung
00:26:09
prohibited the use of gen AI like the
00:26:11
chat GPT on company devices and network
00:26:14
to prevent data laks so again very
00:26:18
important to when you implement AI make
00:26:21
sure that there is a governance and
00:26:22
policy in place if you are using an AI
00:26:26
make sure from time to time to check on
00:26:28
those uh ethics um ethics and principles
00:26:32
that I mentioned earlier transparency
00:26:34
accountability
00:26:36
non-bias um human Centric so let's look
00:26:39
in education how is it applied
00:26:41
personalized learning uh the adoption of
00:26:44
AI and Ed and education really raise
00:26:47
questions about the future role of
00:26:49
universities and faculty but here's how
00:26:53
um how AI application complement rather
00:26:56
than replace the traditional education
00:26:59
structures question there is will we
00:27:02
still need universities or faculty there
00:27:05
are students here right there are
00:27:07
faculty in here so will we still need
00:27:09
University or faculty personalized
00:27:12
learning in AI enhances the learning
00:27:15
experience it adapts the content to
00:27:18
individual student needs and learning
00:27:20
styles so what is the role of the
00:27:23
faculty I'm also an ajun faculty so I'm
00:27:26
saying will I will they still need me so
00:27:29
teachers and professors still play a
00:27:31
crucial role in The Guiding mentoring
00:27:33
and facilitating a deeper understanding
00:27:36
and critical thinking and creativity no
00:27:40
uh elements that AI alone cannot achieve
00:27:44
in terms of automated grading a AI can
00:27:46
automated grading for routine
00:27:48
assignments saving the uh The Faculty
00:27:52
time to focus more on complex student
00:27:54
support so it's more on the personal
00:27:57
relationship and support and mentoring
00:27:59
to the student and the the grading is um
00:28:02
repetitive task something that can be
00:28:04
given to an AI but make sure for those
00:28:08
students when you learn that your grade
00:28:12
is now being automated through AI make
00:28:16
sure that the faculty and I also um call
00:28:19
the attention of the faculty the
00:28:21
decision is still yours and please take
00:28:24
care of reviewing the results of the AI
00:28:27
now the RO Ro of the faculty more again
00:28:29
more on human oversight to ensure the
00:28:32
grating accuracy the fairness and to
00:28:34
provide a personalized feedback
00:28:37
especially for subjective uh assignments
00:28:40
no AI in the workplace it's the um
00:28:44
automation of routine tasks like um here
00:28:48
in the zoom there's already the a
00:28:50
automatic AI uh let's say Note Taker it
00:28:53
creates the transcripts and um it can
00:28:57
actually to some application the slack
00:29:00
it can identify the to-dos or action
00:29:04
items even identifies can depending on
00:29:08
the workflow integration it can even
00:29:10
identify who are the action owners and
00:29:12
the due dates so it will create a
00:29:15
notification that a certain uh action
00:29:17
item discussed and agreed on the meeting
00:29:20
is already due or is assigned to an
00:29:23
individual so um what are the
00:29:26
opportunities for staff in an AI
00:29:28
enhanced uh workplace there are new AI
00:29:32
driven roles AI integration um creates
00:29:36
role like AI trainers or data analysts
00:29:40
or AI Implement uh implementation
00:29:43
Specialists or even ethics officers
00:29:47
so huge or bigger organizations when
00:29:50
they Implement AI they also integrate
00:29:54
new roles like this in their
00:29:55
organization imagine ethics officers is
00:29:57
a part of um an organization now it's a
00:30:01
new role when there is a heavy
00:30:03
implementation of an AI this is where
00:30:06
you will see how they try to um take it
00:30:10
seriously especially in implementing the
00:30:13
ethics and principles so the the focus
00:30:16
now on the individuals or the staff is
00:30:18
on strategic and creative work but uh AI
00:30:23
enabled decision support is there
00:30:26
however the final decision will still uh
00:30:30
lie on the uh individuals the staff most
00:30:33
especially the leadership um skills
00:30:37
development and upskilling this is very
00:30:40
important let let's forget the the our
00:30:44
fear before maybe five years ago or even
00:30:48
more that AI or the robots will replace
00:30:51
us only those who don't know how to use
00:30:53
the AI will be replaced now there are uh
00:30:57
requirements in some um just try looking
00:31:01
at job opportunities on LinkedIn for
00:31:04
example you will find skills and um cap
00:31:08
skills and knowledge that were not there
00:31:10
before even for marketing and social
00:31:13
media try to look at the
00:31:16
requirements what are these requirements
00:31:18
that were not there before imagine
00:31:22
um sometimes they will they are looking
00:31:24
for social media influencer that is now
00:31:27
a new skill that can be uh integrated by
00:31:30
a company when they are looking for
00:31:32
social for marketing and sales um
00:31:34
individuals uh even the use of Tik Tok
00:31:37
no depending on depending on the um
00:31:40
depending on the uh position or the job
00:31:43
that you are applying for look at the
00:31:45
skills that are looking
00:31:49
for uh enabled uh what do you call this
00:31:52
companion working together with an
00:31:56
AI um new graduates my advice to you as
00:32:01
well try to look at use your Linkin
00:32:05
because when you apply for a job the L
00:32:07
the a the Linkin is already has an AI
00:32:10
enabled um uh classification or
00:32:14
qualifying you already how many of your
00:32:16
skills meet the skill requirement of the
00:32:20
job that is being
00:32:22
posted and from there you will know what
00:32:24
are the skills that you need to to
00:32:26
upskill or whether you need to rescale
00:32:29
so preparing for the future to
00:32:31
effectively prepare for the future with
00:32:34
AI individuals need to focus on this
00:32:38
area develop AI
00:32:40
literacy um it involves continuous
00:32:42
learning about the new AI development
00:32:44
the tools and best practices which will
00:32:47
help the uh you as individual to stay
00:32:50
informed and capable of using AI
00:32:52
effectively in your role so very
00:32:55
important it is not asking you to to
00:32:58
trans uh let's say move to another role
00:33:01
but stay informed and how AI can help
00:33:04
you in your or be effective in your role
00:33:08
the ethical awareness it is equally uh
00:33:11
critical staying updated with the AI
00:33:13
ethics the regulations guidelines and
00:33:17
ensure that AI is used responsibly
00:33:20
protecting your use you as a user and
00:33:23
your organization from potential risks
00:33:26
and lastly uh adaptability it is
00:33:29
essential for integrating emerging AI
00:33:32
tools and approaches we cannot get away
00:33:34
with AI anymore and in one conference in
00:33:37
one uh last week in the gitex I was one
00:33:40
of the panelists and one of the audience
00:33:41
asked me how can we stop Ai and I said
00:33:44
why will you stop
00:33:47
AI right and I uh he said that there are
00:33:51
his point of view and highly respected
00:33:53
that is there are now a lot of he he
00:33:57
feels that AI um integration is creating
00:34:02
some not more okay some chaos and it is
00:34:07
a boo um what you call a detrimental
00:34:10
rather than beneficial so I asked him
00:34:14
did you stop
00:34:15
smartphones when you were still using an
00:34:18
noia
00:34:20
3210 okay so were you able to stop the
00:34:24
evolution of smartphone at that time and
00:34:27
and did you ask if you can stop the
00:34:31
development of further or a smartphone
00:34:34
or further enhancing
00:34:36
smartphone so AI is there you have to
00:34:39
embrace it and um you have to live with
00:34:41
it again uh I also do Consulting and at
00:34:46
one point I love giving opportunities I
00:34:50
love mentoring at one point I have to
00:34:53
take someone in the Philippines for a
00:34:56
job for a task
00:34:58
um just also to give an uh an experience
00:35:02
of how it is working in the UAE but
00:35:06
remotely and my question is how good are
00:35:09
you in using chat GPT okay and uh I said
00:35:15
use the chat GPT because I know he
00:35:17
doesn't have the right level of skills
00:35:20
but using the chat GPT will help him
00:35:23
learn what is the subject he has a
00:35:26
certain idea of the subject check the
00:35:28
task but coming out of the uh the
00:35:32
outcome itself I would still prefer
00:35:36
someone who knows how to use GPT if you
00:35:39
don't know it use the GPT you will learn
00:35:42
from it and you will increase your
00:35:44
knowledge from it so don't don't say
00:35:47
that to me if anyone says H CH GPT now
00:35:50
I'm not using my brains no you use your
00:35:52
brains together with the GPT okay I we
00:35:55
still say that in the end we need our
00:35:58
brains to work we need our brains to
00:36:00
work with the AI and lastly uh again I
00:36:05
cannot get away with this slide because
00:36:07
I'm in I am in cyber
00:36:09
security remember that there are cyber
00:36:12
crime in AI cyber crime in AI
00:36:15
highlighting how AI can be exploited by
00:36:19
malicious actors Aid driven social
00:36:22
engineering before we can see we can
00:36:24
determine that there is a malicious
00:36:26
email now with
00:36:28
AI you will barely notice it uh
00:36:32
automated attacks for us we are very
00:36:35
careful about that because AI can
00:36:36
automate cyber attacks like credential
00:36:39
stuffing or distributed denial of
00:36:41
Services Network scanning you know
00:36:44
making them faster and more difficult to
00:36:47
uh to to detect there are defects for
00:36:51
fraud AI generated defects can be used
00:36:54
for financial fraud and in fact there
00:36:56
was um a
00:36:58
it was last February or March that uh a
00:37:02
huge uh company in the
00:37:05
UK W lost 25 million I think double
00:37:10
digit in millions because they approved
00:37:13
the release of such fund and they
00:37:16
thought it was legit because the the um
00:37:20
the adversary even did a zoom meeting
00:37:23
with them and unfortunately it was a
00:37:25
deep fake CEO
00:37:28
AI generated malware the AI helps
00:37:31
attackers identify high value Target
00:37:35
they can predict vulnerabilities and
00:37:38
adopt malware tactics by bypassing the
00:37:42
defenses um AI generated malware can
00:37:46
also create more sophisticated malware
00:37:48
capable of learning from detection
00:37:51
pattern soia already if I am the AI
00:37:53
generated malware I already know that
00:37:56
this is how your AI can detect me
00:37:59
therefore I can also evade your security
00:38:03
measures more effectively uh gathering
00:38:06
information the Cyber criminals use the
00:38:09
AI to to scrape massive amounts of data
00:38:12
from public and private sources um also
00:38:16
reason why I mentioned earlier very
00:38:18
important be be aware about the uh how
00:38:22
your data is being processed whatever in
00:38:25
whatever reason uh you're using an AI
00:38:28
and uh data poisoning so attackers can
00:38:31
feel feed false data into your AI system
00:38:34
to manipulate the outcome or degrade the
00:38:39
performance so in um conclusion there
00:38:43
are potential
00:38:44
benefits but there is an importance of
00:38:48
uh being aware of the security as well
00:38:51
um ethical considerations we have
00:38:54
discussed that there are other principle
00:38:56
AIS ethics and principles be aware of um
00:39:00
other policies globally so you are not
00:39:04
not only informed uh of what is
00:39:06
happening within the country but be uh
00:39:09
let's say um encouraged to learn as well
00:39:13
how the other countries are implementing
00:39:16
it and how you can apply it not maybe
00:39:18
not in your own uh organization but at
00:39:21
least even on a personal level sometimes
00:39:23
we cannot demand more from the company
00:39:26
you know this requires budget this
00:39:29
requires big decision uh but if you can
00:39:32
apply the the uh let's say grab the
00:39:36
benefits of AI for you personally
00:39:39
enhance your well-being and your
00:39:42
profession and your career then do so uh
00:39:45
make sure that you are um preventing an
00:39:49
authorized access if you can um do not
00:39:53
upload so much confidential and personal
00:39:56
in information
00:39:57
in any of the publicly open
00:40:00
Ai and um yeah uh ethical considerations
00:40:06
for for uh work space workplace and in
00:40:10
education maintain fairness prevent
00:40:13
biases uh adhere to privacy regulation
00:40:17
and ethical guidance and in AI
00:40:20
deployment and for those of you who also
00:40:23
would like to know how to implement an
00:40:25
AI governance put governance risk and
00:40:28
compliance in your AI implementation
00:40:31
there is already an AI management system
00:40:33
ISO
00:40:34
4201 I'm in the uh process of take
00:40:37
getting my ISO 420001 lead lead auditor
00:40:41
and there's a lot of things to learn
00:40:43
from this ISO management system
![[19th GSSS Artificial Intelligence dan Blockchain untuk Sustainability]](https://ruhnrawpgxrzdrgaohnf.supabase.co/storage/v1/object/public/images/video/4bZT7ImBtFA/thumbnail.jpg)
