QUESTION AND ANSWER- ARTIFICIAL INTELLIGENCE AFFECTING GLOBAL SECURITY

Parliamentary Intelligence-Security Forum
3 followers
Dec 28, 2025
29TH PARLIAMENTARY INTELLIGENCE-SECURITY FORUM- MADRID, SPAIN🇪🇸
The floor is open for any questions that
you might have. Please introduce
yourself and ask your question. Good
morning everyone. I am Representative
Gerald Inraman from the Philippine House
of Representatives. This question is for
Mr. Jamil Jaffer. I am very much
interested in what you said that uh
innovation should come first before
regulation.
Now in the Philippines, we have a legal
void. We don't have any laws regarding
AI. In fact, we are on the receiving end
of AI that comes from Asian superpowers,
China basically. And we have seen how AI
is being used to manipulate public
opinion also to destabilize our
government. So in our case, we are not
innovators in our country. We haven't
reached that point.
cannot innovation and regulation
start together. To give you an example,
um when we open the market to AI
companies and investors, can we not
limit the type of use that this
technology can be applied to something
like that? I'd like to listen what you
have to say about regulation and
innovation. Thank you. No, thank you for
the question uh congressman and I think
it's a it's a very important point.
There are types of regulation that might
be valuable uh in this domain. Thinking
about where your country and your
companies, your innovators buy their
technology from, right? Thinking about
the capabilities they rely upon, right?
That might be a smart way to think about
regulation. The places to be cautious
about regulation are to think about when
you start limiting the access to data
you might put in there or you start
limiting the access to the compute you
might have or you start limiting access
to the utilization that of that AI data
or that those AI created outputs you
might then have a result that prevents
innovation from happening in those
domains. Now, you might say to yourself,
look, as as the European Union has,
look, it's so important to us that we
protect against harm in certain areas or
we're so worried about discrimination in
certain areas that we're going to rate
the AI usage in that area as high risk
and we're going to subject it to higher
standards. There's nothing wrong with
making that decision. The result of that
decision though often times uh will be
that you tamp down innovation in those
domains, the innovation happens
elsewhere, and then you can utilize it
or not.
Europe has operated that way in a
variety of regimes uh you know for
decades and Europe has thrived
economically relative to the rest of the
world but relative to the United States
when it comes to GDP growth when it
comes to technological innovation when
it comes to creating new capabilities in
this domain for better for worse you're
lagg significantly behind and part of
that not all of that part of that I
think turns on this decision to limit
how you utilize it so smart limits are
okay I think I think you're right to
raise that question. Um, but if you step
too far in an effort to, you know, in in
the with the right mentality, protecting
your citizens, ensuring less harm, um,
you can have these significant knock-on
effects on innovation, and you just have
to make that trade-off. Europe has made
a choice, and we've seen the results.
Yes, sir. You want to comment? Yeah, I
guess my one comment to that is um you
know it might be it might be good to
think about what area of regulation this
is right rather than defining that as an
AI regulation this is really a media
regulation right I think many countries
control ownership of television stations
right radio stations
uh in many ways we've we've created sort
of uh internet media platforms that are
you borderless in the way in which
they're applied. So, it's really not
necessarily a an artificial intelligence
problem. It's about how to media outlets
influence and spread disinformation.
So, I would almost define it that way.
And and I think we've gotten so caught
up in this idea that it's artificial
intelligence that it's manipulating and
causing a disinformation. I think we
should be very clear. It's humans that
are using artificial intelligence tools
across various media outlets to do that,
not artificial intelligence.
I do agree that AI has positive uh uses
and I would also like to comment to what
Mr. Turpin said that in the Philippines
we are now I personally am in a process
of crafting a bill that would actually
regulate social media platforms. We all
value our constitutional right to
freedom of speech and to actually try to
regulate um content creation per se is
something that is not viable and most
probably unconstitutional. So we are
trying to regulate social media
platforms so they can act as gatekeepers
to keep away the purveyors of fake news
etc. So just in the right direction I
think we are headed. Thank you. Yeah.
Yeah. And if I if I can also add, I I I
would look at this from and I think
that's the right approach, Congressman.
I would look at this from the lens of
the activity. So you're not regulating
the technology, but you're regulating
how the technology is being applied.
Right? There are regulations on the
books now. Um whether it's uh say in
financial services, banking prohibiting
discrimination against uh those who are
applying for loans, AI can be used to
discriminate as we know. Um and so those
regulations are on the books that will
prevent that kind of discrimination. But
I think what um professor was saying,
what others have been saying is that
going in with that regulation first
mentality will prevent a lot of those
innovations that will actually be um
again regulated because the activities
are already regulated. you don't need a
whole um new regulation on the
technology but rather on those
activities. So the higher the riskier
those activities are the more say um
consumerf facing people facing um you
want to ensure that there's the
consumers the people are protected the
more regulation the less riskier the
activity the less regulation but there's
a there's a um happy medium or there's a
balance that needs to be. Yeah. Thank
you com we had a question up here.
Yes. Uh thank you very much. Cham
Senator Tito Scoran, chairman of the
foreign relations committee of the
Romanian Senate. Speaker, please
following uh following a little bit your
comments, it's not a question is a
comment but a brief comment. I want to
draw your attention to the fact that
exactly one year ago in May last year,
the Council of Europe, which is the
oldest pan European organization based
in Strasburg, European Court of Human
Rights, standards in democracy, rule of
law, human rights, adopted the first
ever international treaty on artificial
intelligence, but not on artificial
intelligence in relation to what one of
the keynote speakers mentioned, the need
to protect the human dignity and
fundamental human rights. It's a
important uh step forward for a
framework regulation. Um it's not a
perfect one. It's not a perfect legal
solution. Why? Because before the final
adoption by the committee of minister by
the governments of this uh treaty as the
parliamentary assembly of the council of
Europe and I saw a number of colleagues
which are coming from Lithuania as a ban
republic of mold of Romania which are
members of the parliamentary assembly.
We have adopted a positive opinion to
support the adoption of the convention
with a number of amendments and one of
the amendments was related to the need
to adopt a specific regulation against
the abusive uh use of the uh artificial
intelligence that can alter
fundamentally the integrity of the
political electoral processes.
Fortunately, the committee of minister
didn't take on board our amendments for
the sake of a speedy procedure and we
saw 9 months later what happened in my
country in Romania during the
presidential elections. So what I'm
saying is the fact that there is a step
forward. We'll be obliged to update
quite fast in my opinion with a amend
amending protocol of the framework
convention and just to signal the fact
that this is an open international
treaty that can be signed and ratified
by also by the non European states if
there is a specific interest. Thank you.
Thank you for your comment. Yes sir.
Uh yes I I come from Kenya in Africa. Um
the comment that I have in respect of AI
is this.
It's a new technology
but it is created by people with
biasedness.
The question is the bias that is in
included in the AI the algorithm and
other issues. So you can talk about it
because AI does not have a conscience
like a human being. It does not have
that conscience to decide between A and
B. In Africa, we have gone through
colonialism of different types. Now we
are we have come into neoc colonialism.
Now we are coming to AI colonialism
where it depends upon the information
and the data collection. You speak about
socialism in South in in Africa.
Tanzania is not as bad as you say when
you come over. The situation that I'm
trying to possess is this. We are
getting AI colonism in Africa. Whether
is whether it's Congo, whether is South
Africa and other places. We have got
these people who are coming that we had
somebody called the world coin. They
took our um iris
the copy of the iris. Now we don't know
what for. They went to go and develop
something in the United States. So the
the the the classic example that I would
like to try to suggest to you is
regulation. Yes, at your level we wish
to reg we wish to understand it at our
level. What is the African concept of AI
because there's nothing new we were we
were with drums then it came to
communication then we came to computers
and I remember computers were something
that we never knew. When I was a young
man, I used to stand outside a call box
waiting for a telephone call session. I
had arranged one week behold. So I I
think things are changing. Um we in
Africa and I speak for for Kenya, we we
think we are on the forefront of AI but
we are not. The situation is is mis is
misused. So that is when you're talking
about regulation, we're talking about
western standards. I think what we need
to also do is to redesign it to see how
we each country can regulate its AI
according to its own standards. Thank
you very much. Thank you very much.
Further questions. Thank you very much
for the opportunity. Uh
Frederick is my name. I come from the
Parliament of Nigeria particularly the
House of Representatives. Writing on
what my colleague from Kenya has said in
Nigeria, we have tried several times as
a parliament to regulate the social
media. And what Nigerians believe is
that if you regulate anything coming
from the social media, it means that you
are making a law against freedom of
speech.
And so whatever you promote in AI, in
the social media, once you want to
regulate it, it is seen as anti-
people's bill, anti- people's law. And
so Nigerians will come out vehemently
against it and it has always failed. And
with what is coming from the AI,
certainly when you want to attempt to
regulate it, it will be seen as anti-
people's law. And in Africa,
particularly in Nigeria, we haven't
grown to the level where we'll be able
to have tech control of what AI is going
to be delivering to our people. In that
situation, it gives us a lot of concern.
What will be the prescriptions? I think
that's my remark. Thank you.
Thank you very much. Yes, sir. Dr.
Meyer, Paul Mau, president of Montre
College, United States. All three of you
um cited the human factor in AI, the
need for humans to come alongside,
human teaming was one of the commons,
working with humans, the need for
ethical boundaries do no harm. I'd love
to hear one or more of you amplify on
the human factor in AI.
Sure. Um so I just want to be very clear
AI does not operate
without
human guidance. I mean the the type of
artificial intelligence we have now
artificial narrow intelligence is deeply
deeply dependent upon
human teaming. Uh it really only the
applications only work across narrow
tasks.
We hear an awful lot about artificial
general intelligence or artificial super
intelligence things that are in the
future, right? That would operate as a
human operates across multiple, you
know, that has wide competency across
multiple things can operate uh sort of
in a self-aware and a self-functioning
way. But that is simply not what we have
today. Right? Right? So I want to be
very clear that's these are these are
tools that humans use to do jobs, right?
It could be in uh you know creating and
disseminating narratives across social
media, but it's also uh allowing
companies to be able to process large
amounts of their own data to be able to
see faults in manufacturing lines. It's
generally the same things. They're
tools. It's really a continuation of an
information technology revolution that
began in the 1980s. Right? Artificial
intelligence is simply one more piece of
that. It's a human tool, not not its own
thing yet.
Any further comments,
Dr. Cameron?
Thank you so much um Congressman
Pittinger and for the panel uh for your
exemplary uh remarks today. I was
wondering if there is uh any kind of
work ongoing looking at jobs of the
future that AI as was mentioned a minute
ago will start to uh take on some of the
roles that humans have done in their
day-to-day work lives and what will the
work of the future look like and how
will that impact the the contract that
governments have with their citizens in
terms of taxation for instance is is
there any kind of work I think this is
one of the very crucial aspects that we
have to look at
yeah I mean look we we've seen a lot of
changes already um in the way that um
humans interact with AI but also the
role that AI plays in the workforce if
you look at a company like um Anthropic
that is one of the big AI companies they
create they have a model called claude
um I was on a panel with the chief
information security officer of
anthropic and he told the audience that
Today, 30 to 40% of Anthropic's new code
for Claude, its developments on top of
its existing model are done by Claude
itself. The code is actually written by
the by the LLM to improve itself. Um, he
said by the end of the year, 30% today,
by the end of the year, it'll be upwards
of 80%.
They have guardrails in place.
Anthropics is a company that prides
themselves on on responsible innovation.
They have guardrails in place. The
guardrails are something like the
following. I may not have to get this
exactly right, but the way he described
it was if the model improves itself
through this coding methodology by more
than a,000x per year in a single
12-month period, the circuit breaker
kicks in and they'll stop development.
But 999x
in 12 months, that's that's okay. You
know how powerful the models are today.
So, it gives you a sense of some of the
jobs that might might shift away. This
isn't going to be though. While there
may be major shifts, there will be new
opportunities. You've seen already
prompt engineers, people who are
designing queries for AI tasks being
paid upwards of six figures without a
college degree.
So, it's an interesting dynamic in a lot
of ways. Um, there was an initial take
that uh that the advent of generative AI
and its its implications out in the
world um would be that you'd lose a lot
of blue collar jobs. The reality is
actually more likely to have a much more
significant impact on white collar jobs.
Now the upside to that is that as it
replaces these sort of white collar jobs
and and enables everybody to have access
to a legal assistant at their at their
at their fingertips, right? A coder at
their fingertips. Um it'll also take the
people who are good at that stuff and
who are innovative and are creative and
incentivize them to do more and have
more time and more opportunity. So
actually enhance human creativity. I
actually think this is in a lot of ways
the the the angst about AI and its
destruction of the workforce and its
impact on bluecollar workers and its
impact on creativity I think is exactly
the opposite and while it's true that
you will need to ensure that AI models
in order to develop faster have access
to lots of data and we you know in the
US uh we have this doctrine of um fair
use when it comes to copyrighted
materials right there's a debate going
on in the US today if whether the use of
data for AML models is fair use or not.
And my view is if you don't consider the
use of data for to train an AM model of
fair use, you're going to kill the the
baby of AI, you're going to kill the AI
innovation. And so you've got to allow
it to be fair use. Then the question is
how do you ensure that innovators and
creators are incentivized? That's a
different separate matter, right? But I
actually think you're going to see more
creativity, not less with the advent of
generative AI. And so I think there are
really unique opportunities and allowing
people to be more creative. So I'm not
worried about the jobs of the future.
They'll be very different, I think, than
the jobs of today, but they'll be
they'll take advantage of what makes us
uniquely human in things that AI simply
cannot and I I'm not sure ever we'll be
able to do and replace that creativity
and that conscience and the things that
we were uh talking about earlier. And
and to the earlier Jones's question
about from Kenya about um about um the
biases of these models, there's no doubt
that the data you train these models on
particular if it comes from a western
only corpus, right, is going to give it
a certain sensibility. And then if you
if you counsel the model and you
instruct the model to apply say
enlightenment values as I heard one
company has done right that's going to
have a certain flavor to it. You're
exactly right. But nothing prevents a
country in Africa university in Africa
multiple universities coming together
and developing your own model trained on
African data. You leverage other data
from around the globe but trained on
African data train on African values and
developing that. China's done it with
its models. Right? There are challenges
in doing that, right? There are but but
the admin gener actually allows you to
do it faster and faster. So I think
there are real opportunities here um
that that that are are potentially be
taken advantage of. Thank you panelist.
Thank you panelist. Outstanding
presentation.
