Season 1

Energy Security in the CEE Region and AI-Driven Security Solutions - Panel 3- TEB 1st Forum

with Shireen· Austro Papia· 1m

TL;DR

AI's role in energy security faces data and trust challenges.

Synopsis

The panel discusses the intersection of energy security and AI, highlighting the importance of data availability and trust in AI systems. Experts emphasize the need for regulatory frameworks to facilitate AI integration while addressing cybersecurity and geopolitical risks.

Key metrics

by the numbers · 1

15 terawatts
Power in certification queue

Topics

4 tags

AI integrationenergy securitydata availabilityregulatory challenges

Stats

Duration: 35m
Words: 5.3k
Questions: 40

Timeline

5 chapters

Panel Introduction
Panelists introduce the topics of energy security and AI.
AI Status Discussion
Panelists discuss the current status of AI in ensuring energy security.
Trust Issues with AI
Panelists address the trust issues surrounding AI outputs in energy systems.
Regulatory Perspectives
Discussion on the need for regulations to ensure safe AI integration in energy.
Future of AI in Energy
Panelists share expectations for AI's future role in energy security.

Key insights

4 takeaways

01
Data is crucial for AI
AI requires high-quality data to function effectively; poor data leads to poor results.
02
Trust in AI is essential
Building trust in AI outputs is critical for its acceptance in energy systems.
03
Regulatory frameworks needed
Clear regulations are necessary to guide the integration of AI in energy sectors.
04
Cybersecurity risks present
Geopolitical and cybersecurity risks must be managed alongside AI deployment.

Pull quotes

2 quotes

The biggest hurdle would be the trust issue with AI
Kobin
AI needs to be integrated into a lot of different workflows from planning of energy system to operation
Shireen

Transcript793 cuesClick a timestamp to jump

So in the middle exactly
so this is the final final discussion
we're having today. So I see the
networking drinks already prepared. So
bear with me. I'm sure it's going to be
pretty interesting with such a panel.
We are talking today about the two the
combination of the two most pressing and
exciting topics in the energy world
which is NSG security and of course
artificial intelligence.
And these two topics play a very
important role in today's society
today's world where geopolitical and
infrastructural challenges require
tailored and forward-looking solutions.
We already had very insightful
presentations. Thank you for that. So
you already know three of our panelists.
We're quite lucky to get have a fourth
one, Shireen from Osloia.
And just to make it fair, may I kindly
ask you to introduce yourself and
present what you're doing, what your
daily work is. Thank you.
>> thank you so much. thanks also
for having me here today and thank you
to Daniel and Rahul for organizing this
great event actually. yeah so
yeah I'm a senior expert at Ostro
Papia. Austroapia is the association of
the of the paper industry of the
Austrian paper industry and we have
23 members most of them are well
known. So we have like we have Mandi,
we have Hansel, we have ST we have
Mayans. So they are actually global
companies that act in different
parts of the world. yeah and I my
background is also before I joined
Austro Papia in February and before
that I was working for the Austrian
Ministry of Defense and so it's
really interesting to combine
this security approach with energy
and now here I'm glad to join the
panel on energy security and AI.
>> Perfect. Thank you. It's a pleasure
having you here.
>> Thank you so much. So we already had
interesting insightful presentations. I
would like to go a little bit higher
regarding the level talking about energy
security and AI. And there's this
concept of the Gardner hype cycle. I
don't know if you're aware of that. So
it basically says if there's a new
technology, a new innovation,
it gets a trigger then expectations
increase until the peak of ex
exaggerated expectations. So this is
when much too much is expected from this
new innovation. Then there's the draft
of disillusion
where everyone says well it's never
going to fly. And then there's this path
to enlightenment and finally the plateau
of productivity. So one example would be
blockchain technology which
I think and I didn't hear the word
blockchain at all at this event. I'm
sure four five years ago everyone would
be talking about blockchain.
So this is how things change and using
this concept I would be interested in
what is the status of AI
to ensure energy
security today.
Maybe you can also dive into the
different aspects. Where are we
regarding technology,
regulation,
business models and so on? Are there all
the same level? Are we
behind in some areas? Where do we
need the most progress? And if it's
okay, Jesse, I would start with you.
>> Yeah, absolutely. Yeah. So where are we
with AI today?
I think everybody's trying to find their
way around AI right now and
especially regulation and the
end users are trying to find what is
what is the use case of AI and where is
it kind of most prevalent and I would
say where the data is the most
available and accessible is where AI
lives right now and because data
is like the food of AI. You
either if you have bad data, you will
have bad results and that's
that's my view on the topic.
Perfect. Thank you. Let's continue with
sharing. So regarding the data source,
how far is the Austrian industry
regarding data availability? What are
their expectations on AI, on energy
security? What's important for them?
yeah, AI plays a big role and it's
really important. But before answering
this question, I'd like to give a
broader overview on energy security
and what it actually means for the
Austrian industry because it's not only
anymore about energy supply
only but it's also about about profit
profitability and about prices. So
the paper industry is very energy
intensive. So the prices so
actually the main challenge currently
are the rising prices. so
this is this is a very important
aspect and I think AI can also play here
a role in like predicting what
prices can can can be on the
market. But this is this is also
very very important especially
because there are if you compare I
if you compare the prices in
Europe with the US or with China we
have a lot of challenges here so
yeah there there there's a lot to do
and energy security has also a lot to
do with sustainability and I think
this is also something very important to
highlight. So it's not only
about having energy but also
having having sustainable energy.
So renewables it's really important to
to have more renewables and to
think or to think together to bring
together the energy and the
decorization
goals. Actually there are a lot of
regulative issues on the EU level
targeting decarbonization. So the
industry has to decarbonize and
there the AI comes in and this is
something that the Austrian
as the Austro Papier and also the
industries are very much focusing on.
So we have for example research projects
and providing actually AI tools
to how what are actually the best
solutions for the future combining
the least CO2 emissions with
profitable energy prices. So, so yeah,
this is very important.
>> Perfect. Thank you., during your
presentation, you already showed us
quite some impressive demonstration
projects and innovation projects. Where
do you see the application of AI in
terms of energy security? Where what are
the hurdles, the challenges?
>> Yeah. Yeah, regarding
specifically the grid operators, so DSOs
as well as DSOs, they're traditionally
very very conservative. I'm sure you can
>> attest to that. Yeah, very conservative.
So it's very difficult for TSOs, DSOs
across Europe, not just Austria to
adapt adopt new solutions like AI. I
would say the biggest hurdle would be
the trust issue with AI to ensure
that the outputs that the AI can can
produce is actually believable and it
doesn't cause more issues in the system
rather than solving the issues..
>> So that's really the biggest word I
would say and right now AI is of course
being applied to wide variety of domains
across the energy sector and in fact
right now I'm actually working on a
study for the European Commission DJer.
So they contracted us to find out
what are the best use cases for AI. So
that is currently being worked at. So I
would say it's quite broad but there are
of course some specific use cases
especially generating models. a lot
of LMS have been used for inheriting
or is keeping all this this very
important inherent data that that
very senior technicians for example in
DSOS app and when they leave that whole
information is lost. So that's one
interesting use case where you can use
an LLM to train young technicians to be
able to work in this this very
difficult environment. So it's it's
quite broad I would say and of course
right now it's still in definition
phase. There are pilot projects that are
being that are currently running to
test this, but I would say it's still
very early.
>> Perfect. Thank you, Kobin. Impressive
projects you're working on. Where do you
see the main hurdles in getting to
integrate your products? What are the
perspectives of customers regarding AI?
How do you see this?
So to be honest on your last question so
I guess our customers namely the EV
users they do not have so much
expectations towards AI but the thing is
our selling point lies in fully
optimized electricity right and what
we've done already so going on a very
very deep company level what we've done
already is we were doing a lot
of with machine learning, customer data,
etc. to train our models. We've already
built the first APIs to AI
providers. we did not build LLMs yet
on our own, but this is maybe to come in
in the near future as well. so the
first that's the first thing and where I
see on a second level some usage or some
some good
foundation or some good fundament for
using AI is I showed you these all these
buckets which we are where we are
somehow in between as an aggregator
right we have the energy supply we have
the net operator we have the car
manufacturer we have the wall OX
manufacturer and at some point we need
to ensure interoperabilities
in order to be able to
execute the vehicle to grid topic and I
guess in this context AI will support
this wall to somehow make this
interoperability easier and smoother.
>> Perfect. Thank you so much. So Bar you
already touched the topic of trust into
AI models. So I don't know maybe some of
you remember in 2010 there was this
flash crash at the stock market where
automated
algorithms
sent the market into
into chaos for some minutes wiping
out nearly one trillion in value. So
this is one of the first incidents
where automated
algorithms
did not work out as expected.
So how can we ensure that this does not
happen in the energy sector in our
energy models? What is needed to avoid
this? Is this techn from a technologies
perspective from a regulatory
perspective
and basically can we trust AI or do we
have to fear that it will take over the
energy system the world and we are just
passengers on this trip.
>> Yeah, I already I think indicated
this in my my slides. there's not
going to be some kind of a Skynet event
where we hand over the entire power grid
or the entire energy system to an AI
that controls everything on its own.
that's not going to happen. I think it's
always a co-pilot and so we will take
the help of AI to understand to get some
deep insights into certain aspects but
it's always humanentric. Human always is
in the loop.
So I would like to assure everybody
that's not the direction that we're
taking at all. Yeah. There's not going
to be some kind of fully autonomous
AI controlling all aspects of our
system.
>> Would you like to add to that? Yeah.
>> Yeah. Also, been working with
utilities the past 10 years and seeing
how kind of cautious they are about new
technology into the grid. This is not
going to happen. the
piloting and then kind of digging into
deep into the technology that they're
deploying it will never happen that they
just say okay this is a nice AI just
take over that will never happen.
>> Perfect. Thank you.
Do you want to add something?
>> I fully agree.
>> Fully agree. They will not take the
world over.
>> Perfect.
So, Sheree, from the perspective of an
industrial energy consumer, where do you
see the greatest opportunities,
challenges, risks in ensuring energy
security for
for industrial consumers and for the
papers industry in particular?
>> yeah. thank you for this important
question. And I think it's important to
have this broader picture. So to
combine industrial competitiveness
together with decarbonization and
energy. So I think it's important to
have this broad dimension. and
here AI can play can play a very
important role in showing
what path industry can can actually
take. but also of course there are
a lot of risks to that. so
coming back to energy security we have a
lot of geopolitical risks. and then
we have also linked to that we
have also cyber security risks as well.
And this is of course something that
is also on the on the top on
the agenda on the top of the agenda
of a also the Austrian paper
industries. So when having AI
in any systems it's always about
looking into the into the tools into
the models and it's also it's really
important that the provider of an
software actually tells exactly where
the points are where the security
measures are and yeah and I think
this this is this is really important.
maybe also again I'd like
also again to highlight this
decarbonization issue because there
are a lot of regul regulative
regulatories on EU level and this
is this is something that AI already
plays a role and I think this
this this is also something that will
actually develop more and more in the
into the in the future. So
>> perfect.
>> I hope I answered your question.
>> Perfect. Thank you so much. Kobin, we
already touched upon the topic of data
as food for AI models.
Where do you see the biggest hurdles
in getting data? So, and you're
combining different data sources from
different
companies, organizations.
Which are the ones who already are there
in providing the necessary data? Where
do you see gaps? What what needs to be
done to provide the data in a format
that's usable for your models?
Difficult question.
so right now it's we so two parts right
what we're doing now in the company
in that context we are very very
interdependent on data and customer
behavior of our customers that means we
are dependent on the data of our EV
users this is basically easy to get
because we are generating the data and
we are setting up the data in the way as
we want it to be set up, right?
once you think one step ahead of that
and coming back to vehicle to grid topic
I guess it is so we have the
experts here. So if we will build the
intersection to grid operators and to
the energy market I get I guess this
data is not so difficult to get either
right it depends so it's how you how
it's organized within the local
>> yeah it depends really because right
now at least in Austria so data
protection is very very strong and to
get the customer data I don't know the
smart meter data or any other data
related to consumption production This
is very difficult. Yeah. So we have the
the law local energy community that
could be one option to get this. But
again like you mentioned electric
vehicle is not defined under this law.
So perhaps there's a some kind of a link
there so that you are part of a
community but this this electric vehicle
V2G is kind of an add-on but that's not
defined on the mall right now.
and that's the
additional thing I guess the most
complicated part in terms of getting
data and in terms of somehow ensuring
interoperabilities
in that context is the EV
manufacturer
because I guess these kind of
stakeholders need some regulatory
pressure to somehow release data to
somehow participate and this is not
defined at this point of
But we are full of hope because I
was reading through the program of our
new government and it was a very very
strong and bold bullet point that
they integrating everything which is
required for vehicle to grid in the
future laws namely EV for example.
>> Yeah this is true I can attest to that.
Yeah, it is right now sitting on the
desk of a colleague of mine for public
inputs right now at this stage and
and yeah, I think the again this is
not fully clear but battery is defined
or is going to be defined and if battery
is defined then it's not a it's not far
away to define electric vehicle which is
which is a battery which moves.
>> Yeah.
>> Yeah. The LVG is a law that is also
important for the for the Austrian paper
industry because it will also like
enable more flexibility of energy
of use of energy and this is
really important. however it's not
yeah there are still some challenges
also coming with this with this
law but yeah and batteries are included.
So of course batteries are also aspect
very important aspect also for the for
the industry.
>> Perfect. Thank you for that.
>> Jesse, what would you add regarding data
availability, data formats,
standardization of data input, people
willing to provide the necessary data
and so on.
>> Yeah, people probably are
available and willing to give the
data, but the utilities for example are
are not. they don't like their data
to be outside of their fence. So to
be able to deploy AI and what we have
learned is that we need to deploy the AI
inside of the fence and bring onpremise
solutions for the utilities so they are
easily implemented there. They don't
have to go through regulatory hurdles of
of sending data outside of their fences
and things like that. So that's
that's then derisking the innovation
for the utilities for example.
>> Perfect. Shireen
could you add from a consu industrial
consumer perspective any insight in
the willingness to share data to give
data outside? Would you rather prefer to
have as Jesse said the models running
inside
the companies?
yeah I companies I agree with you
companies don't don't really like
sharing data of course we
had for example a big research project
on how to decarbonize and which
which also which solutions for the
paper industry are viable for the
future with also scenarios and things
like that. and yeah it was it was
a research project that was
financed by some of the companies and
and there were some discussions on
will they like provide it for
everyone or not and of course it's
something when it comes to data it's
it's always yeah it's always hard
actually.
Yeah, but regulators can Yeah, but the
regulators can force utilities and DSOs
to provide this data. Especially in
Austria where EG
local law is forcing DSOS to provide
this data to all the customers.
So it has to be done through the
regulation way
>> regulator
this is something where I have to say
that's that's of course not not
interesting because everything that you
know comes by forcing It's more
problematic I would say. So
>> I think you need both.
>> Let me let me let me try let me try to
rephrase that bit. So I didn't use the
right terminology. So I should have said
the customers are willing willing to
participate in the energy community and
to enable this they need to they need
their data and the DSOs are owning this
data and that's the reason they need to
provide this data. So it's for the
benefit of the customer. So let me
rephrase that again. It's it's a bit of
a difference here because DSOs are
are fiduciary public entities of course
privately owned but they're for the
public but industries of course
industries are slightly different. Yeah.
So it has
>> your own yeah industrial or business
goals that they have and I can
completely understand that.
>> Yeah. And of course it's also about the
competition. yeah you cannot
share everything with other companies
but since you are you are
focusing more on energy communities
because this is also something that is
being discussed currently but very
theoretically how in future
baby industry can like have a own energy
community to share energy within
within the certain industry baby
industries but this is very
theoretically the advantage would be
the network costs that might be not
there anymore and then the prices would
be better. But yeah right now there
is no like legislative like like a
law actually that is making that
possible. But who is the
producer in that scenario? Because it
basically should balance out producers
and consumers and paper industry is
strongly consuming. Right.
>> Exactly. It's very strongly consuming
but it's also many companies also
produce electricity. and so yeah it's
it's they are also producing and
there biogas for example is very
important within the within the paper
industry yeah
>> but I think it's currently but within
the current regulations of local
contributors it is possible for
industries to join the communities
this is actually currently possible. But
again, of course, it's not fully open.
There are some restrictions on that. So,
it's not the same as some individual
household customers join, but industries
and companies can join right now.
Yeah. With some restrictions, of course.
>> Perfect. Thank you. Unfortunately, we
are somehow running out of time. So, I
would like to go through one last round.
And we started the panel discussing
about the status quo. I would like now
like to move towards the future. So what
are your expectations for 10 minutes?
yeah, but it's an important topic for
what's going to happen in the future and
what are the most exciting developments
or breakthroughs you foresee in the
integration of AI and for energy
security
and how do we ensure that AI does not
only improve system efficiency but
creates a more sustainable, resilient
and equitable energy future.
Who wants to start? Any volunteer or do
I have to pick one?
So in the future for AI, the
the big hurdle will or or the big jump
will come once quantum computing
comes in and enables a lot more
computing power that we have right now.
Then the AI models are on steroids
basically. and that's kind of a
scary place also because then like no
password and nothing is safe anymore in
the in the internet or the grid or
anywhere. So there's definitely needed
some good regulation to kind of mitigate
the risks but I think the upside is
also quite exciting of infinite
possibilities of utilizing data for
for different AI models.
>> So summarizing you would say the
bottleneck currently for the AI models
is computing power.
>> No no not right now. it will
just enable a next jump. right now
the data and availability of the data
is probably the bigger bigger hurdle and
and
yeah and also like we have mentioned
trust of the AI. AI is something new.
It's it's it's like the new big
efficiency engine that just came about
and people are not really trusting
it that oh can it really deliver all
these good things and u yeah that
that will be kind of the bigger hurdle
there but yeah
>> and what could we do to increase this
trust in into AI?
explanability goes a long way
I think and kind of showcasing
what data is used and how. and
yeah I think if if your buddy
tells about an awesome fishing trip and
and tells you that I got caught a
two 2 m fish you won't believe him. But
if he shows a picture then you believe
him right. So it's about the
explanability of a story and where
wherever AI is implemented.
>> Perfect. Thank you. So Shireen from your
side do you see that it's already well
explained how AI works the industry?
What are the expectations of the
industry for the future using AI?
of course it is a AI is considered as
the most important thing that will
have a very will play a very
big role and already now AI is more
and more used. So not only in the energy
sector but also in other you
know when when producing paper
for example here AI already plays a
big role. So definitely and there
are a lot of discussions and debates
on that. we are also part of the of
TP which is the confederation of the
European paper industries and there
always sections and sections and
panels and meetings on AI in the
paper industry related also to energy
very much to decarbonization. So I
would say it is consider considered
as the solution for many things.
but yeah still we have some
regulation regulative issues to
think about and there are many risks of
course cyber security as I have
mentioned but I would I
wouldn't say that there is no trust
actually maybe maybe the trust maybe
there is not no not trust to everyone
let's say you have other other
experiences of course but I think I
think there is a trust in AI as a
whole.
>> Yeah. I see a lot of exciting
things. I'm I'm hopeful. I want to be
hopeful. I want to be open-minded. I
want to be early adopter of the
solutions that are being developed
at the moment. Yeah, of course, trust.
Trust and explanability is kind of
very important. but I would say that
there's a lot of research that is being
done on this exact issue across the
world. So of course not specific to
energy industry but AI in general. So
explanability is kind of the most
important criteria right now across the
world for adoption. and yeah in
the study that I mentioned earlier for
the European Commission we are
trying to also be very hopeful in this
study and try to recommend that AI needs
to be integrated into a lot of different
workflows from planning of energy system
to operation of energy system and to
ultimately commissioning of energy
system. all of these things, the whole
spectrum. we want to see more and
more AI applications in the whole
ecosystem of energy and I think there
are some some severe bottlenecks right
now in energy transition. For example,
grid connection and certification
bottleneck right now. So I think 15
terowatts of power is just sitting in
the queue right now in across Europe
because there are not enough electrical
engineers to certify plants to be
connected so they can start selling
power to the to the TSO or to the DSO
and these kind of issues can be
eliminated or at least reduced with the
help of AI. I think it's a very powerful
tool. We should be of course cautious
about use of AI but we need to be also
optimistic and hopeful that it's going
to deliver some very unique as well as
solutions that are not typically
out there in the in the system
realm.
>> Perfect. Thank you. So Copinian,
how do you see the future for vehicles
to grid? What are currently the hurdles?
How does the a perfect world in a few
decades look like from from your
perspective?
>> So I spoke so much about grit. Now I
would like to add one holistic thing and
and connecting what you have said. I
guess the whole energy system as we
know it today is highly interdependent,
right? We have PV systems, they're
somehow connected to batteries, to cars
in the very near future. We have the
grid operators, we have the energy
suppliers, and they're all somehow
swinging and talking with each other.
And this is super complex obviously. And
I guess AI could be the trigger point or
the mean to make this somehow smoothly
going. This is I guess the big chance of
AI and this is a very holistic view. I
I'm afraid a very future fear because
right now as said we do not even have a
clear regulatory understanding whether
we are allowed to push energy from an
out from a car battery back into the
grid or not. So these are highly
regulatory aspects and I do believe as
well that at some point you need some
you need to have some regulatory
pressure and some regulatory
forcement or however you would like to
call this at least you need very
dedicated laws to that and I think this
is the big bottleneck right now.
>> Perfect. And do you think regulators are
on track setting up all these
regulations or
>> I believe so. I strongly believe so. But
you must imagine what happened in
the last 10 years. Our perception, our
understanding of what is the primary
energy source on this earth has totally
changed. Right? 10 years ago we would
have said there are some fossil things.
Right? Right now we say it's electricity
and end of story. And to adjust all our
laws to that is difficult and takes some
time.
>> Perfect. Thank you so much. I believe
this was quite an engaging and
insightful discussion. So thank you for
that. It made my life much easier that
you were so proactive contributing.
So I would like to summarize that when
we thoughtfully apply AI, we can enhance
our forecasting efficiency and
responsiveness in our energy systems.
And while it might introduce new
complexities and vulnerabilities, I
believe we're pretty good on track on
making the future energy system much
more resilient, much more efficient, and
and basically a better system for the
future. So thank you for that and with
this I would
>> last keynote.
>> Last keynote
>> enough to summarize.
>> Sure. Do you want to add something on
that?
>> Do you have any questions?
>> Yeah. Let's go for the audience for some
questions.
>> Question.
>> Everyone is waiting for the drinks.
>> Yeah.
>> Okay. So a big applause for the last
Thank you. Thank

Energy Security in the CEE Region and AI-Driven Security Solutions - Panel 3- TEB 1st Forum

Data is crucial for AI

Trust in AI is essential

Regulatory frameworks needed

Cybersecurity risks present