July 12, 2022 • 17 mins
In this episode, Ariana Smetana (Digital Transformation Expert), Noelle Silver (AI Ethics & Education Leader), Tolani Jaiye-Tikolo (Intelligent Automation Expert), and Andreas Welsch discuss building trust in Artificial Intelligence (AI) & automation. Ariana, Noelle, and Tolani share key insights on why keeping humans in the loop is a critical factor for trust in AI and provide valuable examples for listeners looking to introduce AI in their business.
Key topics:
- Start transformation project right
- Build trust in AI with one simple action
- Create automation balance with humans ‘in the loop’
Listen to the full episode to hear how you can:
- Look at business value and business objectives first
- Deliver on your automation promise to build trust
- Keep a human in the loop to improve your data
Watch this episode on YouTube: https://youtu.be/dHWICJ7l878
Questions or suggestions? Send me a Text Message.
***********
Disclaimer: Views are the participants’ own and do not represent those of any participant’s past, present, or future employers. Participation in this event is independent of any potential business relationship (past, present, or future) between the participants or between their employers.
Level up your AI Leadership game with the AI Leadership Handbook:
https://www.aileadershiphandbook.com
More details:
https://www.intelligence-briefing.com
All episodes:
https://www.intelligence-briefing.com/podcast
Get a weekly thought-provoking post in your inbox:
https://www.intelligence-briefing.com/newsletter
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Andreas Welsch (00:00):
Today we have a special episode for you.
I've been playing one of myfavorite retro games again,
What's the BUZZ?
And I'm really stuck.
And so that's where I've askedthree experts to help me on the
quest for trust in AI.
Ariana, Tolani, and Noelle.
Thank you so much for joining.
Noelle Silver (00:22):
Thanks so much.
We're excited to be here.
Andreas Welsch (00:25):
Awesome.
Why don't you tell us a littlebit about yourself?
Ariana Smetana (00:30):
Thank you, Andres for having me today.
I'm Ariana Smetana and I'm thefounder and CEO of the AccelIQ
digital innovation consultancy.
And we guide companies todifferent creative
differentiating value for theircustomers, employees with
development of digital tools,solutions, and advanced
(00:54):
technologies that can implementin their organizations.
Over the last three years, I'vebeen focusing predominantly on
emerging technologies, and Ihave a very keen interest in AI
and I use applications andactually creating the business
value the most importantly with.
I came to AI field a rathernon-traditional route.
(01:15):
I'm not technologist.
I'm actually economist and havea business degree and long
history of business knowledgeand actually being an
entrepreneur.
So that's where I come from andbring value to the businesses
with use of technology.
Andreas Welsch (01:34):
Fantastic.
Thank you so much for joining.
I'm so glad you're you're hereand to have you on.
Likewise.
Yeah.
Over to Noelle.
Noelle Silver (01:43):
Awesome.
Hello.
I'm so excited about this game.
It seems so fun.
I hope I can help.
I don't know.
But I'm Noelle Silver.
I'm the CEO and founder of AILeadership Institute, where we
educate executives across theglobe around artificial
intelligence and really aroundresponsible AI at scale.
(02:03):
One of the other things thoughthat I do on a daily basis is
I'm an executive at IBM and Ialways think it's interesting
cuz my career 22 years agostarted at IBM.
A very different IBM but it'sinteresting, we're gonna talk
about like transformation and AIand building inclusive teams.
And all of these things I'vewatched happen in my career.
(02:25):
So super excited to hear fromall of the people playing the
game today and happy to be apart of it.
And of course you can alwaysfind me on LinkedIn if you guys
need another LinkedIn friend.
Andreas Welsch (02:35):
And that's how we met as well.
So I'm really excited about thepower of the platform and
connecting on that topic thatwe're so passionate about.
So then maybe over to Tolani.
Great to have you on as well.
Tolani Jaiye-Tikolo (02:46):
Thank you.
I'm so super excited to be here.
Cuz we've been talking onLinkedIn for god knows how long.
It's really, it's sometimes it'slike a dream for me.
Just as you've mentioned, myname's Tolani Jaiye-Tikolo.
I'm the founder of RPA JargonBuster.
The good thing about RPA JargonBuster, it's not just, busting
(03:10):
jargons or busting memes.
It's actually a small businessbased out of Ireland.
And what we do is quite simple.
We do four things very well.
We do intelligent automationresearch.
Adversary consulting andcorporate trainings but doesn't,
literally about myself.
I've been in the digitaltransformation space for seven
(03:31):
years or I've been more focusedaround robotic process
automation, and that's where Iliterally started my career
from.
And within that know timeframeI've seen, RPA evolve into other
sort of other technologies suchas what we are beginning to talk
about like intelligentautomation.
And that's really why I foundedmy own business to really, get
(03:56):
into that market share and seewhat's going on in that space.
So what we do is crossconsortium rem.
We work with SMEs and mid-marketfinancial services and
healthcare.
It's a no-brainer why I'm in thefinancial services.
I'm also an employee as well.
I work with Allied Irish Bank asan RPA lead.
(04:16):
But I'm really happy to be here.
Thanks for having me..
Andreas Welsch (04:20):
Excellent.
Thank you so much.
I don't think I realized it whenI set it up and the fact that
all three of you are founders ofyour own companies that's
exciting too.
So really glad that you're ableto add that perspective to it as
well.
So maybe just a quick shout outto folks in the audience.
(04:41):
If you're just joining thestream, drop a comment in the
chat, what tricks you're lookingfor, but like last time, no
cheating allowed.
What do you say?
Should we start playing?
Ariana Smetana (04:54):
Absolutely, yes.
Let's do it.
Tolani Jaiye-Tikolo (04:56):
Yeah, let's do it.
Andreas Welsch (04:57):
Excellent.
So this one is a warmup.
When I hit the buzzer, you'llsee a sentence and I really need
your help.
So can you answer with the firstthing that comes to mind and
why, in your own words.
And so together you have 60seconds for your answer.
Keep it short, right?
So team, are you ready forWhat's the BUZZ?
Noelle Silver (05:20):
Yes.
Andreas Welsch (05:22):
Let's do this.
If AI were an animal, what wouldit be?
Maybe Noelle.
What do you think
Noelle Silver (05:32):
I'm like it has to be a unicorn.
Because some people believe init and some people don't.
But it's capable of incrediblethings, if you believe.
So I like unicorns or rhinos,which are actually the real life
version of a unicorn if you needa real animal.
(05:54):
Cuz some people say unicornsaren't real.
Who else got one?
Andreas Welsch (05:58):
Perfect.
Who else?
Ariana Smetana (06:03):
It's funny, first thing came to mind was
actually a monkey and I think,it's really constantly working
and dealing with lots of dataand information and you're not
coming out with the answers.
Like you said, some peoplebelieve it, some don't believe
it, and it thinks something.
Some people think it's a monkeybusiness, but it's definitely
(06:23):
not
Andreas Welsch (06:25):
Lovely.
I love it.
How about you, Tolani?
I know we're out of time, butkeep going.
Tolani Jaiye-Tikolo (06:30):
Yeah, so I think it would be a chameleon.
It's a lizzard just becausepeople think they know it, but
they don't.
So it always changes based onwhat you fitted and how it turns
to work around differentenvironments.
So I would say chameleon.
Andreas Welsch (06:48):
I like that.
Fantastic.
Thank you so much.
Thanks for coming up with thosethings on the fly.
I really appreciate it.
Ariana Smetana (06:57):
Fantastic.
Noelle Silver (06:58):
Like I went first.
That's fair.
Andreas Welsch (07:00):
I put you right on the spot.
Noelle Silver (07:01):
I know.
Andreas Welsch (07:02):
I think that part around it's different
things.
It changes.
That's something I feel reallyresonates at least with me and
I'm sure with folks in theaudiences as well.
And changes and means so manydifferent things to so many
different people.
Let's maybe take a look at thefirst question and what that is.
And that is (07:25):
What's the first thing in a transformation?
So wondering Ariana if you canhelp me with that.
What's the first thing you do ina transformation?
Ariana Smetana (07:37):
So what I advise my clients to first think of is
business problem.
We started with the businessproblem.
What is the business problem andhow can it be solved with AI?
Should it be solved with AI?
And why is this problemimportant for the organization
to be solved?
There are so many tools that weuse in organizations and AI is
(08:02):
another tool to be applied.
Certainly it's business problem.
The next one will be businessvalue of solving that problem
and how we going to use that togenerate the value and then
creating what is the datastrategy objective functions.
Modeling and all of that comeslater.
But certainly, knowing ourresources from the data, from
(08:23):
the people, from the processes,these are so many layers that
need to be addressed.
But I would say, businessproblem is the first one to
start with.
Andreas Welsch (08:33):
Fantastic.
And I think that's a theme wekeep hearing, right?
Start with the business problem.
Worry about technology later.
If you do it the other wayaround you're just in love with
technology and you struggle tomake that case.
Ariana Smetana (08:50):
The business value is where the every company
will want to use that.
Otherwise, there are so manyother tools you can use for
solving problems and decisionmaking.
Andreas Welsch (09:01):
Let's maybe then take a look at question number
two that we have here.
And that is (09:08):
How do you build trust in AI?
That kind of goes along with it.
If you start with the businessproblem and you roll it out to
people, how do you make surethat they actually appreciate
what you're doing here and thatthey accept what you're doing?
And I know Noelle, you've beendoing quite a lot of work in
that area.
(09:28):
What do you recommend?
Noelle Silver (09:29):
Absolutely, this is a good question.
And it's so interesting, like ofcourse it's level two cuz the
first thing, as you mentioned inany digital transformation,
right Ariana, is figuring outwhat problem are we gonna solve?
And then you think, okay, howcan I solve this, like the
fastest way I can with the besttechnology?
And AI is often thrown on thetable as a tool.
(09:52):
And the operative word is trust.
And so the way that I buildtrust when I go into an
organization and I'm talkingabout the dream of AI which I am
very good about selling thedream of what you can do with
all this cool technology, buttrust only comes when I actually
can deliver on the dream,deliver on the promise.
And it ties directly what towhat you said, right?
(10:13):
As soon as an organizationculturally understands what
problem they're gonna solve andwhat tangible business outcome
they want, that's becomesliterally the way I build trust
in AI.
I build an AI system thatdelivers on that value.
So oftentimes like simple,intelligent workflows, right?
If somebody has a manualprocess, I can go in and just
(10:35):
say, okay, I commit to improvingyour throughput of customers by
35% with the use of AI.
Then, my job is to actuallydeliver on that promise, on that
35%, and as soon as I do, trustis built, but it's built in
these like little smallprojects.
Gartner even said like over thelast five years, 90% of projects
(10:56):
failed that had the word AI inthem.
And the reason why was exactlywhat Ariana said, that, we'd
never identified the exactproblem we wanted to solve, nor
did we attach an actual value tothat solution.
And that's where AI is the mostpowerful.
And so yeah, I'd say we buildtrust by delivering results on
those business outcomes bybuilding models that solve those
(11:18):
very specific.
Andreas Welsch (11:21):
Fantastic.
I really love that.
And make it tangible.
And put your money where yourmouth is.
With all the hypes still aroundAI and the AI whitewashing, it's
it's so easy to claim there's AIin everything.
And all the promises that wemake, we better keep them.
Noelle Silver (11:39):
That's right.
Andreas Welsch (11:39):
Especially if you wanna be credible and win
trust in AI.
Excellent.
So then maybe let's take a lookat the next question here, and
that is question number three: Why do businesses still want a (11:48):
undefined
human in the loop?
And maybe that's somethingTotani that you can maybe help
me with.
I know you work a lot on RPA andintelligent document processing,
but there is a lot of peoplestill in the loop.
Why do businesses still want tohuman it?
Tolani Jaiye-Tikolo (12:08):
I think I'll look at you from two
different angles.
And I'm probably gonna take youfrom what Noelle stopped.
When we talk about businessoutcome and delivery on those
business outcome, we still needto advise people that AI would
never be a hundred percent interms of predictions.
So people would need to be waryof this and also acknowledge
(12:28):
that with AI you might have 80,90%, but you still need human in
the loop to help refine the viewof the world.
And I'm probably gonna give agood example here particularly
around if you have, let's say achild or if you have a nephew or
a niece.
So whatever the case might be.
You take them to the zoo andthey identify different animals.
(12:52):
This is it's a lion, this istiger, or wherever the case
might be.
Then tomorrow I come back and Itell the child I give them a
picture of a fox and I tell themit's a dog.
The first thing they're gonna dois argue with you and say, no,
this is not a dog.
Right?
There are two things are gonnahappen.
Either I get convinced by thatchild that it's not a dog, or I
(13:13):
convince that child that it's adog.
And let's say, because of who Iam, I convince the child that
it's a dog.
What's happened is the view ofhow they see things is
completely different.
Every time they see a fox, itbecomes a dog.
Every time they say a dog, it'sstill a dog.
And that's why I tell them thatAI a hundred percent, you don't
get that a hundred percent.
(13:33):
But you keep changing the courseof AI by refining, putting our
human input into the system.
And I think the second part ofit is, again, I might probably
just put it on the businessesbecause we can.
Businesses fail to find gooddata sometimes.
What I would always advise is,you want to build, let's say a
(13:56):
good model.
You need for documents, youprobably need about 500 for good
prediction, I tell you that.
How many people in theorganization can go into
wherever they have it, intotheir lockers or wherever on the
network, and give me 500different samples.
It's not gonna happen.
But what we can do with humanlook is they keep in as much as
(14:18):
we have a small sample, we cankeep refining that over time.
And AI becomes fresher than howit was yesterday.
And this is where humans reallychange that course.
But given the ai, thetransparency it needs and the
oversight it needs and it keepsimproving over time, just not to
take too much of your time,probably just stay on those two
(14:38):
examples.
Andreas Welsch (14:39):
Thanks.
The part around having thathuman in the loops even if you
do have the data, you still needto label it and you need someone
to tell you, what are maybe thebounding boxes around some of
the text and all the manual andeven still laborious tasks while
you're trying to get rid of somelaborious tasks.
That's where people can help youalready in the process today.
(15:01):
If you augment it.
If you make that step part ofthe process, build a good data
set that you can then do your AIwith.
Perfect.
Thank you so much.
So that's awesome.
Team, we did it.
Thank you so much for playingWhat's the BUZZ?
today! So let me summarize realquick.
(15:21):
First of all, if you think aboutdigital transformation, make
sure you look at business valueand business objectives first.
That's where it all shouldstart.
Technology comes second.
Then secondly, while you'rebuilding that talk track and
getting that buy-in, whetherit's from your customers or from
your internal stakeholders, makesure that you also deliver on
(15:42):
the promise to keep thecredibility to build that trust
and reinforce it.
And one way to do that is tohave a human in the loop for
processes, especially wherethey're document based or where
you can and should reviewsomething to make sure that your
data set is built and getsbetter over time.
Fantastic.
So we're getting close to theend of the show today.
(16:04):
Thank you so much for joining usand for sharing your experience
with us.
I really appreciate it.
Ariana Smetana (16:10):
Thank you.
Noelle Silver (16:10):
Thank you.
It was so fun.
Today we have a special episode for you.
I've been playing one of myfavorite retro games again,
What's the BUZZ?
And I'm really stuck.
And so that's where I've askedthree experts to help me on the
quest for trust in AI.
Ariana, Tolani, and Noelle.
Thank you so much for joining.
Noelle Silver (00:22):
Thanks so much.
We're excited to be here.
Andreas Welsch (00:25):
Awesome.
Why don't you tell us a littlebit about yourself?
Ariana Smetana (00:30):
Thank you, Andres for having me today.
I'm Ariana Smetana and I'm thefounder and CEO of the AccelIQ
digital innovation consultancy.
And we guide companies todifferent creative
differentiating value for theircustomers, employees with
development of digital tools,solutions, and advanced
(00:54):
technologies that can implementin their organizations.
Over the last three years, I'vebeen focusing predominantly on
emerging technologies, and Ihave a very keen interest in AI
and I use applications andactually creating the business
value the most importantly with.
I came to AI field a rathernon-traditional route.
(01:15):
I'm not technologist.
I'm actually economist and havea business degree and long
history of business knowledgeand actually being an
entrepreneur.
So that's where I come from andbring value to the businesses
with use of technology.
Andreas Welsch (01:34):
Fantastic.
Thank you so much for joining.
I'm so glad you're you're hereand to have you on.
Likewise.
Yeah.
Over to Noelle.
Noelle Silver (01:43):
Awesome.
Hello.
I'm so excited about this game.
It seems so fun.
I hope I can help.
I don't know.
But I'm Noelle Silver.
I'm the CEO and founder of AILeadership Institute, where we
educate executives across theglobe around artificial
intelligence and really aroundresponsible AI at scale.
(02:03):
One of the other things thoughthat I do on a daily basis is
I'm an executive at IBM and Ialways think it's interesting
cuz my career 22 years agostarted at IBM.
A very different IBM but it'sinteresting, we're gonna talk
about like transformation and AIand building inclusive teams.
And all of these things I'vewatched happen in my career.
(02:25):
So super excited to hear fromall of the people playing the
game today and happy to be apart of it.
And of course you can alwaysfind me on LinkedIn if you guys
need another LinkedIn friend.
Andreas Welsch (02:35):
And that's how we met as well.
So I'm really excited about thepower of the platform and
connecting on that topic thatwe're so passionate about.
So then maybe over to Tolani.
Great to have you on as well.
Tolani Jaiye-Tikolo (02:46):
Thank you.
I'm so super excited to be here.
Cuz we've been talking onLinkedIn for god knows how long.
It's really, it's sometimes it'slike a dream for me.
Just as you've mentioned, myname's Tolani Jaiye-Tikolo.
I'm the founder of RPA JargonBuster.
The good thing about RPA JargonBuster, it's not just, busting
(03:10):
jargons or busting memes.
It's actually a small businessbased out of Ireland.
And what we do is quite simple.
We do four things very well.
We do intelligent automationresearch.
Adversary consulting andcorporate trainings but doesn't,
literally about myself.
I've been in the digitaltransformation space for seven
(03:31):
years or I've been more focusedaround robotic process
automation, and that's where Iliterally started my career
from.
And within that know timeframeI've seen, RPA evolve into other
sort of other technologies suchas what we are beginning to talk
about like intelligentautomation.
And that's really why I foundedmy own business to really, get
(03:56):
into that market share and seewhat's going on in that space.
So what we do is crossconsortium rem.
We work with SMEs and mid-marketfinancial services and
healthcare.
It's a no-brainer why I'm in thefinancial services.
I'm also an employee as well.
I work with Allied Irish Bank asan RPA lead.
(04:16):
But I'm really happy to be here.
Thanks for having me..
Andreas Welsch (04:20):
Excellent.
Thank you so much.
I don't think I realized it whenI set it up and the fact that
all three of you are founders ofyour own companies that's
exciting too.
So really glad that you're ableto add that perspective to it as
well.
So maybe just a quick shout outto folks in the audience.
(04:41):
If you're just joining thestream, drop a comment in the
chat, what tricks you're lookingfor, but like last time, no
cheating allowed.
What do you say?
Should we start playing?
Ariana Smetana (04:54):
Absolutely, yes.
Let's do it.
Tolani Jaiye-Tikolo (04:56):
Yeah, let's do it.
Andreas Welsch (04:57):
Excellent.
So this one is a warmup.
When I hit the buzzer, you'llsee a sentence and I really need
your help.
So can you answer with the firstthing that comes to mind and
why, in your own words.
And so together you have 60seconds for your answer.
Keep it short, right?
So team, are you ready forWhat's the BUZZ?
Noelle Silver (05:20):
Yes.
Andreas Welsch (05:22):
Let's do this.
If AI were an animal, what wouldit be?
Maybe Noelle.
What do you think
Noelle Silver (05:32):
I'm like it has to be a unicorn.
Because some people believe init and some people don't.
But it's capable of incrediblethings, if you believe.
So I like unicorns or rhinos,which are actually the real life
version of a unicorn if you needa real animal.
(05:54):
Cuz some people say unicornsaren't real.
Who else got one?
Andreas Welsch (05:58):
Perfect.
Who else?
Ariana Smetana (06:03):
It's funny, first thing came to mind was
actually a monkey and I think,it's really constantly working
and dealing with lots of dataand information and you're not
coming out with the answers.
Like you said, some peoplebelieve it, some don't believe
it, and it thinks something.
Some people think it's a monkeybusiness, but it's definitely
(06:23):
not
Andreas Welsch (06:25):
Lovely.
I love it.
How about you, Tolani?
I know we're out of time, butkeep going.
Tolani Jaiye-Tikolo (06:30):
Yeah, so I think it would be a chameleon.
It's a lizzard just becausepeople think they know it, but
they don't.
So it always changes based onwhat you fitted and how it turns
to work around differentenvironments.
So I would say chameleon.
Andreas Welsch (06:48):
I like that.
Fantastic.
Thank you so much.
Thanks for coming up with thosethings on the fly.
I really appreciate it.
Ariana Smetana (06:57):
Fantastic.
Noelle Silver (06:58):
Like I went first.
That's fair.
Andreas Welsch (07:00):
I put you right on the spot.
Noelle Silver (07:01):
I know.
Andreas Welsch (07:02):
I think that part around it's different
things.
It changes.
That's something I feel reallyresonates at least with me and
I'm sure with folks in theaudiences as well.
And changes and means so manydifferent things to so many
different people.
Let's maybe take a look at thefirst question and what that is.
And that is (07:25):
What's the first thing in a transformation?
So wondering Ariana if you canhelp me with that.
What's the first thing you do ina transformation?
Ariana Smetana (07:37):
So what I advise my clients to first think of is
business problem.
We started with the businessproblem.
What is the business problem andhow can it be solved with AI?
Should it be solved with AI?
And why is this problemimportant for the organization
to be solved?
There are so many tools that weuse in organizations and AI is
(08:02):
another tool to be applied.
Certainly it's business problem.
The next one will be businessvalue of solving that problem
and how we going to use that togenerate the value and then
creating what is the datastrategy objective functions.
Modeling and all of that comeslater.
But certainly, knowing ourresources from the data, from
(08:23):
the people, from the processes,these are so many layers that
need to be addressed.
But I would say, businessproblem is the first one to
start with.
Andreas Welsch (08:33):
Fantastic.
And I think that's a theme wekeep hearing, right?
Start with the business problem.
Worry about technology later.
If you do it the other wayaround you're just in love with
technology and you struggle tomake that case.
Ariana Smetana (08:50):
The business value is where the every company
will want to use that.
Otherwise, there are so manyother tools you can use for
solving problems and decisionmaking.
Andreas Welsch (09:01):
Let's maybe then take a look at question number
two that we have here.
And that is (09:08):
How do you build trust in AI?
That kind of goes along with it.
If you start with the businessproblem and you roll it out to
people, how do you make surethat they actually appreciate
what you're doing here and thatthey accept what you're doing?
And I know Noelle, you've beendoing quite a lot of work in
that area.
(09:28):
What do you recommend?
Noelle Silver (09:29):
Absolutely, this is a good question.
And it's so interesting, like ofcourse it's level two cuz the
first thing, as you mentioned inany digital transformation,
right Ariana, is figuring outwhat problem are we gonna solve?
And then you think, okay, howcan I solve this, like the
fastest way I can with the besttechnology?
And AI is often thrown on thetable as a tool.
(09:52):
And the operative word is trust.
And so the way that I buildtrust when I go into an
organization and I'm talkingabout the dream of AI which I am
very good about selling thedream of what you can do with
all this cool technology, buttrust only comes when I actually
can deliver on the dream,deliver on the promise.
And it ties directly what towhat you said, right?
(10:13):
As soon as an organizationculturally understands what
problem they're gonna solve andwhat tangible business outcome
they want, that's becomesliterally the way I build trust
in AI.
I build an AI system thatdelivers on that value.
So oftentimes like simple,intelligent workflows, right?
If somebody has a manualprocess, I can go in and just
(10:35):
say, okay, I commit to improvingyour throughput of customers by
35% with the use of AI.
Then, my job is to actuallydeliver on that promise, on that
35%, and as soon as I do, trustis built, but it's built in
these like little smallprojects.
Gartner even said like over thelast five years, 90% of projects
(10:56):
failed that had the word AI inthem.
And the reason why was exactlywhat Ariana said, that, we'd
never identified the exactproblem we wanted to solve, nor
did we attach an actual value tothat solution.
And that's where AI is the mostpowerful.
And so yeah, I'd say we buildtrust by delivering results on
those business outcomes bybuilding models that solve those
(11:18):
very specific.
Andreas Welsch (11:21):
Fantastic.
I really love that.
And make it tangible.
And put your money where yourmouth is.
With all the hypes still aroundAI and the AI whitewashing, it's
it's so easy to claim there's AIin everything.
And all the promises that wemake, we better keep them.
Noelle Silver (11:39):
That's right.
Andreas Welsch (11:39):
Especially if you wanna be credible and win
trust in AI.
Excellent.
So then maybe let's take a lookat the next question here, and
that is question number three: Why do businesses still want a (11:48):
undefined
human in the loop?
And maybe that's somethingTotani that you can maybe help
me with.
I know you work a lot on RPA andintelligent document processing,
but there is a lot of peoplestill in the loop.
Why do businesses still want tohuman it?
Tolani Jaiye-Tikolo (12:08):
I think I'll look at you from two
different angles.
And I'm probably gonna take youfrom what Noelle stopped.
When we talk about businessoutcome and delivery on those
business outcome, we still needto advise people that AI would
never be a hundred percent interms of predictions.
So people would need to be waryof this and also acknowledge
(12:28):
that with AI you might have 80,90%, but you still need human in
the loop to help refine the viewof the world.
And I'm probably gonna give agood example here particularly
around if you have, let's say achild or if you have a nephew or
a niece.
So whatever the case might be.
You take them to the zoo andthey identify different animals.
(12:52):
This is it's a lion, this istiger, or wherever the case
might be.
Then tomorrow I come back and Itell the child I give them a
picture of a fox and I tell themit's a dog.
The first thing they're gonna dois argue with you and say, no,
this is not a dog.
Right?
There are two things are gonnahappen.
Either I get convinced by thatchild that it's not a dog, or I
(13:13):
convince that child that it's adog.
And let's say, because of who Iam, I convince the child that
it's a dog.
What's happened is the view ofhow they see things is
completely different.
Every time they see a fox, itbecomes a dog.
Every time they say a dog, it'sstill a dog.
And that's why I tell them thatAI a hundred percent, you don't
get that a hundred percent.
(13:33):
But you keep changing the courseof AI by refining, putting our
human input into the system.
And I think the second part ofit is, again, I might probably
just put it on the businessesbecause we can.
Businesses fail to find gooddata sometimes.
What I would always advise is,you want to build, let's say a
(13:56):
good model.
You need for documents, youprobably need about 500 for good
prediction, I tell you that.
How many people in theorganization can go into
wherever they have it, intotheir lockers or wherever on the
network, and give me 500different samples.
It's not gonna happen.
But what we can do with humanlook is they keep in as much as
(14:18):
we have a small sample, we cankeep refining that over time.
And AI becomes fresher than howit was yesterday.
And this is where humans reallychange that course.
But given the ai, thetransparency it needs and the
oversight it needs and it keepsimproving over time, just not to
take too much of your time,probably just stay on those two
(14:38):
examples.
Andreas Welsch (14:39):
Thanks.
The part around having thathuman in the loops even if you
do have the data, you still needto label it and you need someone
to tell you, what are maybe thebounding boxes around some of
the text and all the manual andeven still laborious tasks while
you're trying to get rid of somelaborious tasks.
That's where people can help youalready in the process today.
(15:01):
If you augment it.
If you make that step part ofthe process, build a good data
set that you can then do your AIwith.
Perfect.
Thank you so much.
So that's awesome.
Team, we did it.
Thank you so much for playingWhat's the BUZZ?
today! So let me summarize realquick.
(15:21):
First of all, if you think aboutdigital transformation, make
sure you look at business valueand business objectives first.
That's where it all shouldstart.
Technology comes second.
Then secondly, while you'rebuilding that talk track and
getting that buy-in, whetherit's from your customers or from
your internal stakeholders, makesure that you also deliver on
(15:42):
the promise to keep thecredibility to build that trust
and reinforce it.
And one way to do that is tohave a human in the loop for
processes, especially wherethey're document based or where
you can and should reviewsomething to make sure that your
data set is built and getsbetter over time.
Fantastic.
So we're getting close to theend of the show today.
(16:04):
Thank you so much for joining usand for sharing your experience
with us.
I really appreciate it.
Ariana Smetana (16:10):
Thank you.
Noelle Silver (16:10):
Thank you.
It was so fun.
Popular Podcasts
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com
24/7 News: The Latest
The latest news in 4 minutes updated every hour, every day.
Therapy Gecko
An unlicensed lizard psychologist travels the universe talking to strangers about absolutely nothing. TO CALL THE GECKO: follow me on https://www.twitch.tv/lyleforever to get a notification for when I am taking calls. I am usually live Mondays, Wednesdays, and Fridays but lately a lot of other times too. I am a gecko.