"Maintaining An Open Source Toolkit" with Bellingcat Staff and Toolkit Guardians

WEBVTT
00:00:03.950 --> 00:00:06.990
You're listening to a stage talk titled Maintaining
00:00:06.990 --> 00:00:09.849
an Open Source Toolkit. This week we were joined
00:00:09.849 --> 00:00:12.769
by Bellingcat's Johanna Wild and Bellingcat volunteer
00:00:12.769 --> 00:00:16.030
community members, Anisa Sabir and Martin Sona.
00:00:16.589 --> 00:00:19.129
They are all, alongside others, instrumental
00:00:19.129 --> 00:00:21.570
to the build and maintenance of the Bellingcat
00:00:21.570 --> 00:00:24.670
Online Open Source Investigations Toolkit, a
00:00:24.670 --> 00:00:27.829
resource used by over two million people. looking
00:00:27.829 --> 00:00:30.489
for reliable information on open source tools.
00:00:30.949 --> 00:00:32.990
They spent the hour discussing how they make
00:00:32.990 --> 00:00:35.570
decisions on the toolkit and keep the list of
00:00:35.570 --> 00:00:38.149
applications up to date. You can find links to
00:00:38.149 --> 00:00:40.270
all the resources mentioned in the talk in the
00:00:40.270 --> 00:00:42.649
podcast description. This talk was hosted by
00:00:42.649 --> 00:00:44.969
me, Charlotte Marr on Thursday the 8th of October,
00:00:45.030 --> 00:00:55.619
2026, in the Bell and Cat Discord server. Okay,
00:00:56.380 --> 00:01:00.420
I reckon we should get going. We have a lot to
00:01:00.420 --> 00:01:03.460
talk about today. So welcome everyone to this
00:01:03.460 --> 00:01:05.840
week's State Talk and thanks for joining us.
00:01:06.159 --> 00:01:08.299
Today we're going behind the scenes of a resource
00:01:08.299 --> 00:01:11.379
that many people here may be familiar with or
00:01:11.379 --> 00:01:14.560
could find useful, the Bellingcat online open
00:01:14.560 --> 00:01:18.099
source investigations toolkit. The toolkit looks
00:01:18.099 --> 00:01:21.140
at first glance like a collection of useful links
00:01:21.140 --> 00:01:23.980
and tools, but maintaining a resource like this
00:01:23.980 --> 00:01:27.340
at scale raises some much bigger questions. The
00:01:27.340 --> 00:01:29.840
open source investigation landscape is constantly
00:01:29.840 --> 00:01:33.120
changing. New tools appear, existing ones disappear
00:01:33.120 --> 00:01:35.799
or change their functionality. Platforms change
00:01:35.799 --> 00:01:38.500
their policies and technical architecture and
00:01:38.500 --> 00:01:41.359
techniques that work today may not work tomorrow.
00:01:41.879 --> 00:01:43.980
Looking at many of the tools that we used to
00:01:43.980 --> 00:01:46.760
use for Facebook, for example. So there's an
00:01:46.760 --> 00:01:49.299
interesting editorial and methodological challenge
00:01:49.299 --> 00:01:51.780
here. How do you decide which tools are worth
00:01:51.780 --> 00:01:54.359
recommending? How do you test them? And how do
00:01:54.359 --> 00:01:56.819
you assess their reliability and limitations?
00:01:57.400 --> 00:01:59.560
How do you provide guidance that is genuinely
00:01:59.560 --> 00:02:02.659
useful without giving users a false sense of
00:02:02.659 --> 00:02:05.040
certainty? That's what we're going to explore
00:02:05.040 --> 00:02:07.719
today, not just the toolkit itself, but the work,
00:02:07.900 --> 00:02:10.539
judgment and investigation that goes into maintaining
00:02:10.539 --> 00:02:14.099
a resource used by millions of people. We'll
00:02:14.099 --> 00:02:16.400
hear from the people behind the toolkit about
00:02:16.400 --> 00:02:19.020
how they discover and evaluate tools how they
00:02:19.020 --> 00:02:20.840
deal with change, and what they've learned from
00:02:20.840 --> 00:02:23.939
maintaining it over the past two years. Let me
00:02:23.939 --> 00:02:25.960
introduce you to my colleague, Johanna Wild,
00:02:26.120 --> 00:02:28.099
who oversees the development of the toolkit,
00:02:28.360 --> 00:02:30.780
and with her are two members of our Bellingcat
00:02:30.780 --> 00:02:33.560
volunteer community who are also toolkit guardians,
00:02:34.020 --> 00:02:37.319
Anissa Sabir and Martin Sona. Before they walk
00:02:37.319 --> 00:02:39.419
us through their work, a short reminder that
00:02:39.419 --> 00:02:42.240
we will have plenty of time to answer your questions.
00:02:42.479 --> 00:02:44.919
And you can place your questions in the chat
00:02:44.919 --> 00:02:47.259
accessible on the right hand corner of your screen.
00:02:47.400 --> 00:02:49.900
But please remember that this is being audio
00:02:49.900 --> 00:02:52.539
recorded for the podcast. So if you don't want
00:02:52.539 --> 00:02:55.080
me to measure your username, please add that
00:02:55.080 --> 00:02:59.280
as a note in your question. Okay, Johanna, I
00:02:59.280 --> 00:03:03.590
will leave the stage to you. Thank you. I'm so
00:03:03.590 --> 00:03:07.560
happy to be here. Hi, everyone. I start by giving
00:03:07.560 --> 00:03:10.599
you a little bit of background information of
00:03:10.599 --> 00:03:14.800
what our toolkit is, because I assume not everyone
00:03:14.800 --> 00:03:18.280
here knows exactly what it is or has used it.
00:03:18.620 --> 00:03:21.879
Hopefully many, but probably not everyone. And
00:03:21.879 --> 00:03:25.560
here I start. So the reason why we are here today
00:03:25.560 --> 00:03:28.919
is because our toolkit just turned two years
00:03:28.919 --> 00:03:33.810
old, which is really, really amazing. More than
00:03:33.810 --> 00:03:39.169
two years ago, I had the vague idea that we need
00:03:39.169 --> 00:03:42.069
a new type of toolkit because Bellingcat had
00:03:42.069 --> 00:03:46.110
a toolkit before, but it was basically a sheet
00:03:46.110 --> 00:03:49.729
where we listed tools and it was super helpful.
00:03:50.229 --> 00:03:53.770
We could see that people use it and come to it
00:03:53.770 --> 00:03:57.949
quite regularly, but we internally honestly didn't
00:03:57.949 --> 00:04:02.020
manage to keep it updated. I felt sometimes a
00:04:02.020 --> 00:04:04.419
little bit ashamed because this sheet was online
00:04:04.419 --> 00:04:07.080
and we just didn't manage to update it because
00:04:07.080 --> 00:04:10.520
it takes work. And it still does for this version
00:04:10.520 --> 00:04:14.580
too, but I knew we needed something else because
00:04:14.580 --> 00:04:18.680
people use toolkits and they are searching for
00:04:18.680 --> 00:04:22.639
information about tools. And this is how I had
00:04:22.639 --> 00:04:27.050
came up with this idea that if we want to build
00:04:27.050 --> 00:04:30.649
and create a toolkit that makes sense to everyone.
00:04:30.689 --> 00:04:33.389
We need to do it in a collaborative way and in
00:04:33.389 --> 00:04:36.329
a way more collaborative way than we had done
00:04:36.329 --> 00:04:40.029
it before. We had done it mainly within the Bellingcat
00:04:40.029 --> 00:04:43.230
staff team. So I knew we need to open it up.
00:04:43.269 --> 00:04:46.689
And this is how we started this. In my point
00:04:46.689 --> 00:04:50.490
of view, still, I'm very impressed what into
00:04:50.490 --> 00:04:54.689
what it developed. This toolkit to which many
00:04:54.689 --> 00:04:57.250
people in our volunteer community contribute.
00:04:57.620 --> 00:05:00.040
And I just want to talk you through a little
00:05:00.040 --> 00:05:04.639
bit how do we work on this toolkit, right? How
00:05:04.639 --> 00:05:07.300
does it work in the background? But let me first
00:05:07.300 --> 00:05:12.459
of all show to some of you the toolkit and how
00:05:12.459 --> 00:05:15.399
it looks like. So if you cannot see the toolkit
00:05:15.399 --> 00:05:18.720
right now, you can imagine it as just a page
00:05:18.720 --> 00:05:22.899
with different tool categories like geolocation
00:05:22.899 --> 00:05:25.720
maps and satellites, people, websites, and so
00:05:25.720 --> 00:05:29.459
on. And in all those categories, you can click
00:05:29.459 --> 00:05:33.779
on them and you find a list of tools. But what
00:05:33.779 --> 00:05:38.220
is super important for us is that it's not only
00:05:38.220 --> 00:05:41.339
a list of tools. So that was super important
00:05:41.339 --> 00:05:46.600
to me personally, because there are so many great
00:05:46.600 --> 00:05:50.360
toolkit and other types of resources out there
00:05:50.360 --> 00:05:55.290
that list tools, and they are really fast. including
00:05:55.290 --> 00:05:59.189
the newest tools probably on a daily basis, but
00:05:59.189 --> 00:06:05.649
quite regularly. If you know such toolkits, please
00:06:05.649 --> 00:06:08.430
share with us. I'm always collecting interesting
00:06:08.430 --> 00:06:11.769
new toolkits and I like them, but our toolkit
00:06:11.769 --> 00:06:14.509
is different in that sense. It's not meant to
00:06:14.509 --> 00:06:18.529
be the fastest. It's not meant to include the
00:06:18.529 --> 00:06:21.730
newest tools. we do something different. And
00:06:21.730 --> 00:06:24.790
what we really try to do is we take the time
00:06:24.790 --> 00:06:28.769
to look at individual tools and to really test
00:06:28.769 --> 00:06:32.050
them. So if you look at our toolkit, you will
00:06:32.050 --> 00:06:36.490
see there is always a link. And this link leads
00:06:36.490 --> 00:06:41.589
to a page. And on this page, we have quite extensive
00:06:41.589 --> 00:06:44.879
text about the tool. And this is the result of
00:06:44.879 --> 00:06:47.639
the work of our volunteers who basically test
00:06:47.639 --> 00:06:49.800
the tools and then they write what they find.
00:06:50.199 --> 00:06:52.860
And I feel this is the special thing about our
00:06:52.860 --> 00:06:55.800
specific resource, while there are many other
00:06:55.800 --> 00:06:58.620
great resources out there, it's really that we
00:06:58.620 --> 00:07:02.360
test everything in depth and fair the results,
00:07:02.540 --> 00:07:06.480
right? Because this is what sometimes has been
00:07:06.480 --> 00:07:10.500
missing in our field. We like to list a lot of
00:07:10.500 --> 00:07:14.079
tools and we often then already recommend them
00:07:14.079 --> 00:07:17.060
to others, but sometimes we haven't even yet
00:07:17.060 --> 00:07:19.319
tested them. And some tools honestly are just
00:07:19.319 --> 00:07:22.480
not good, or they're a little bit shady, or they're
00:07:22.480 --> 00:07:26.579
only good for one specific thing, but not for
00:07:26.579 --> 00:07:29.519
another thing. And that's why we really wanted
00:07:29.519 --> 00:07:34.259
to look deep into what does a specific tool do,
00:07:34.259 --> 00:07:38.180
and what does it do specifically for open source
00:07:38.180 --> 00:07:41.439
investigations, because Some of the tools we
00:07:41.439 --> 00:07:44.879
use, they have not been built specifically for
00:07:44.879 --> 00:07:47.379
open source researchers in mind, right? They
00:07:47.379 --> 00:07:50.240
have been built with other target groups in mind.
00:07:51.139 --> 00:07:55.459
I always give the example here, for instance,
00:07:55.819 --> 00:07:59.720
peak visor is a tool that has been built for
00:07:59.720 --> 00:08:02.439
people who maybe want to go into the mountains.
00:08:02.980 --> 00:08:06.620
And we now use it for geolocation purposes. And
00:08:06.620 --> 00:08:09.439
there are many, many other cases like that. And
00:08:09.439 --> 00:08:12.519
that means that we really have to look at which
00:08:12.519 --> 00:08:15.379
features of the tool, what exactly of a specific
00:08:15.379 --> 00:08:19.459
tool can be used and are useful for open source
00:08:19.459 --> 00:08:22.600
researchers and what other features are maybe
00:08:22.600 --> 00:08:28.160
not as useful. And yeah, so also what I find
00:08:28.160 --> 00:08:32.460
is special and it also makes our whole project
00:08:32.460 --> 00:08:37.279
quite time intense is that we constantly monitor
00:08:37.279 --> 00:08:41.580
what happens with a specific tool, because tools
00:08:41.580 --> 00:08:45.639
change all the time. Toolkits get outdated honestly
00:08:45.639 --> 00:08:50.080
very quickly if you don't look at them constantly
00:08:50.080 --> 00:08:52.860
and update them constantly. So that's why we
00:08:52.860 --> 00:08:57.360
really have a very rigorous update rhythm. So
00:08:57.360 --> 00:09:02.090
every other month, our contributors look at the
00:09:02.090 --> 00:09:05.110
tools and see what changed and then they changed
00:09:05.110 --> 00:09:08.009
what they wrote about the tool. So that makes
00:09:08.009 --> 00:09:12.110
it really, really resource and time intense for
00:09:12.110 --> 00:09:16.110
all of us, but it also became very clear that
00:09:16.110 --> 00:09:18.990
this is the real value of our toolkit, right?
00:09:19.049 --> 00:09:24.370
So it's really very, very useful for researchers
00:09:24.370 --> 00:09:27.740
to know, okay, I don't have to use this tool
00:09:27.740 --> 00:09:30.639
because things have already changed. So maybe
00:09:30.639 --> 00:09:34.320
I can go to another tool that seems more useful.
00:09:35.299 --> 00:09:40.980
And if we look at the text, the tool reviews,
00:09:41.279 --> 00:09:44.659
as we often call them, that we have in our toolkit,
00:09:45.399 --> 00:09:48.759
they have a very specific structure. And this
00:09:48.759 --> 00:09:52.340
structure really tries to capture the things
00:09:52.340 --> 00:09:54.659
that are important for open source researchers.
00:09:55.289 --> 00:09:59.750
They usually start with a description of, yeah,
00:09:59.750 --> 00:10:03.370
how does this tool work? And tips, how can we
00:10:03.370 --> 00:10:07.769
use this tool? Where do I need to click? Often
00:10:07.769 --> 00:10:11.669
things like that. But then we also have sections
00:10:11.669 --> 00:10:15.429
like, is this tool free to use, right? Because
00:10:15.429 --> 00:10:18.789
that is super important for many people in the
00:10:18.789 --> 00:10:21.970
open source researcher community. Not everyone
00:10:21.970 --> 00:10:26.700
has budgets for tool licenses. Then also how
00:10:26.700 --> 00:10:30.299
difficult is the tool to use? Because I need
00:10:30.299 --> 00:10:33.480
to know, do I need 10 hours to learn how to use
00:10:33.480 --> 00:10:37.659
the tool or can I just go in and use the tool
00:10:37.659 --> 00:10:41.240
right away? And what we have learned is super
00:10:41.240 --> 00:10:44.519
important. And that's also one of the sections
00:10:44.519 --> 00:10:48.519
that we have started to focus on more and more
00:10:48.519 --> 00:10:51.759
is it's very important to describe the limitations
00:10:51.759 --> 00:10:55.840
of the tool after testing it because that helps
00:10:55.840 --> 00:11:00.440
others avoid using a tool that is useless. And
00:11:00.440 --> 00:11:03.500
equally important are ethical considerations.
00:11:03.559 --> 00:11:07.600
We really analyze, okay, what type of ethical
00:11:07.600 --> 00:11:11.220
challenges would come up for someone who uses
00:11:11.220 --> 00:11:15.019
the tool? And yeah, how could we think about,
00:11:15.100 --> 00:11:18.940
how could we approach that? And yeah, the structure
00:11:18.940 --> 00:11:22.139
of our toolkit then also continues. We provide
00:11:22.139 --> 00:11:26.210
guides and articles if available. We also give
00:11:26.210 --> 00:11:28.769
information about the tool provider, specifically
00:11:28.769 --> 00:11:31.649
in which country is the tool provider based.
00:11:32.889 --> 00:11:35.809
And something that we really also learned over
00:11:35.809 --> 00:11:39.610
the years and our toolkit has become now two
00:11:39.610 --> 00:11:43.509
years old, is that it's super important for us
00:11:43.509 --> 00:11:47.610
to show similar tools for every tool that we
00:11:47.610 --> 00:11:52.190
list. And if we can, if there are similar tools,
00:11:52.629 --> 00:11:56.259
we also try to show, okay, and what other similar
00:11:56.259 --> 00:12:00.879
tools could a researcher also use on top of the
00:12:00.879 --> 00:12:08.000
tool that they have already on hand. I also want
00:12:08.000 --> 00:12:12.139
to show you a little bit the background process,
00:12:12.519 --> 00:12:15.080
because this is something that is usually not
00:12:15.080 --> 00:12:18.879
so visible. All the things we do in the background.
00:12:19.220 --> 00:12:22.860
So if you go to a toolkit page, it mainly looks
00:12:22.860 --> 00:12:26.159
like a page, right? But in the background a lot
00:12:26.159 --> 00:12:29.840
is going on and I have tried to capture the process
00:12:29.840 --> 00:12:34.100
shortly how it looks like. So we have a group
00:12:34.100 --> 00:12:37.860
of volunteers who regularly test these tools
00:12:37.860 --> 00:12:42.480
and write the tool reviews and once they did
00:12:42.480 --> 00:12:46.840
that they submit a draft in a system that we
00:12:46.840 --> 00:12:49.679
have for the toolkit. And once this draft is
00:12:49.679 --> 00:12:53.139
submitted, Bellingcat staff, at the moment it's
00:12:53.139 --> 00:12:56.419
in most cases me, we are looking at the draft
00:12:56.419 --> 00:13:00.980
and might discuss with the person who wrote the
00:13:00.980 --> 00:13:05.000
draft what we maybe might change, maybe we might
00:13:05.000 --> 00:13:07.980
want to add some things, maybe not. And once
00:13:07.980 --> 00:13:10.720
this process has happened, only then we publish
00:13:10.720 --> 00:13:13.399
it in the toolkit. So there is really a process
00:13:13.399 --> 00:13:17.120
behind that you cannot see. but it takes a lot
00:13:17.120 --> 00:13:19.840
of time and that is also super important because
00:13:19.840 --> 00:13:24.980
we want to make only quality full reviews public
00:13:24.980 --> 00:13:28.299
of course and then what happens is that every
00:13:28.299 --> 00:13:31.559
other month the people contributing to the toolkit
00:13:31.559 --> 00:13:34.720
they check again what is online and check is
00:13:34.720 --> 00:13:37.399
it still accurate what we have published and
00:13:37.399 --> 00:13:40.139
this is also where our toolkit guardians come
00:13:40.139 --> 00:13:44.990
in and two of them are here today they help every
00:13:44.990 --> 00:13:48.330
contributor every other month with these updates.
00:13:48.649 --> 00:13:51.809
So in the background, they are supporting not
00:13:51.809 --> 00:13:55.250
only their own tool reviews, but they are basically
00:13:55.250 --> 00:13:58.169
supporting the whole team behind the toolkit
00:13:58.169 --> 00:14:03.289
by helping others to improve, write their tool
00:14:03.289 --> 00:14:06.370
draft and help them maybe sometimes to test their
00:14:06.370 --> 00:14:08.129
tools because they have a lot of experience.
00:14:08.570 --> 00:14:10.669
The process in the background, which I wanted
00:14:10.669 --> 00:14:12.870
to share with you just because usually I think
00:14:12.870 --> 00:14:19.700
that is not visible. And before I hand over to
00:14:19.700 --> 00:14:23.940
our great guardians in a minute, I just wanted
00:14:23.940 --> 00:14:28.179
to show you why we are still here. So the reason
00:14:28.179 --> 00:14:31.080
why we put so much effort also at Bellingcat
00:14:31.080 --> 00:14:34.019
into this toolkit together with our volunteer
00:14:34.019 --> 00:14:37.500
community is really because it works. So the
00:14:37.500 --> 00:14:44.879
toolkit went online in September 24, And now
00:14:44.879 --> 00:14:49.659
we are two years in, and we have had over two
00:14:49.659 --> 00:14:53.519
million page views in this time, even a bit more,
00:14:53.659 --> 00:14:57.840
but we lack a little bit of tracking data. And
00:14:57.840 --> 00:15:01.200
it's very consistent. Many people come back again
00:15:01.200 --> 00:15:04.720
after they use the toolkit once. So it's slowly
00:15:04.720 --> 00:15:08.259
going up. More and more people are using it.
00:15:08.399 --> 00:15:12.220
And maybe just that's always quite interesting.
00:15:12.460 --> 00:15:14.659
We have really users from all around the world.
00:15:14.820 --> 00:15:19.440
So of course, from the US, UK, but amongst the
00:15:19.440 --> 00:15:23.659
top three countries with our most biggest groups
00:15:23.659 --> 00:15:26.679
of users, it's India, for instance, and then
00:15:26.679 --> 00:15:30.039
various European countries, but also Brazil.
00:15:30.179 --> 00:15:32.980
So it's very mixed, which is super, super amazing.
00:15:33.620 --> 00:15:37.220
And my last insight, maybe for me, it's always
00:15:37.220 --> 00:15:41.389
super interesting to see what type of tools people
00:15:41.389 --> 00:15:45.730
like most. So they clearly like to look most
00:15:45.730 --> 00:15:49.950
at simple tools that you can use very quickly
00:15:49.950 --> 00:15:53.789
without learning much. And over the two years,
00:15:54.070 --> 00:15:58.429
very consistently people have always liked most
00:15:58.429 --> 00:16:04.429
tools that help you find usernames across different
00:16:04.429 --> 00:16:10.419
social networks or platforms. or a tool that
00:16:10.419 --> 00:16:15.399
helps you find public telegram content. One tool,
00:16:15.639 --> 00:16:19.259
TeleGargo, is quite up in the list. And then
00:16:19.259 --> 00:16:23.139
username tools like Blackbird, these tools have
00:16:23.139 --> 00:16:25.799
been most successful. But if we only look at
00:16:25.799 --> 00:16:29.779
our own smaller community, then anything related
00:16:29.779 --> 00:16:33.039
to geolocation comes up a bit more. Yeah, that
00:16:33.039 --> 00:16:37.200
was just a short overview. Thank you so much.
00:16:37.940 --> 00:16:40.779
And yeah, you can also donate to the toolkit,
00:16:41.080 --> 00:16:45.620
I just mentioned it, because the resources that
00:16:45.620 --> 00:16:49.700
go in, they can be a lot. So we put a lot of
00:16:49.700 --> 00:16:54.539
effort in there. And it goes into working together.
00:16:54.909 --> 00:16:57.970
and in managing the process in the background.
00:16:58.149 --> 00:17:00.370
So this is what we're really focusing on from
00:17:00.370 --> 00:17:05.589
the Bellingcat perspective as a team of staff
00:17:05.589 --> 00:17:10.130
members. But the main work is done by our volunteers
00:17:10.130 --> 00:17:13.410
and also the two guardians that are there today.
00:17:14.789 --> 00:17:17.009
Hi everyone and thank you for having me today.
00:17:17.029 --> 00:17:21.190
It's great to be here. So I oversee the image
00:17:21.190 --> 00:17:24.470
video category of the toolkit and this is the
00:17:24.470 --> 00:17:27.220
place that can get you started with the right
00:17:27.220 --> 00:17:32.759
tools in your visual investigation. And I would
00:17:32.759 --> 00:17:36.839
broadly divide this tool, this area into three
00:17:36.839 --> 00:17:40.039
main categories depending on your research needs
00:17:40.039 --> 00:17:42.539
and they may be different from researcher to
00:17:42.539 --> 00:17:45.839
researcher. So number one, if there's a video
00:17:45.839 --> 00:17:48.619
fresh out of a conflict zone and you need to
00:17:48.619 --> 00:17:50.880
geolocate it, the first thing you would need
00:17:50.880 --> 00:17:55.579
to do is get the keyframes of the video and then
00:17:55.579 --> 00:17:58.079
run reverse image searches for that video. So
00:17:58.079 --> 00:18:00.839
we just have the right tool to get you started
00:18:00.839 --> 00:18:03.160
with it, which is called like Invid Verification
00:18:03.160 --> 00:18:07.500
Plugin. And this is how the dashboard of the
00:18:07.500 --> 00:18:10.299
tool looks like. And from here, then you can
00:18:10.299 --> 00:18:13.079
run reverse image searches using various search
00:18:13.079 --> 00:18:18.619
engines. Secondly, if the target of your investigation
00:18:18.619 --> 00:18:21.319
is a certain person like a bad guy that you want
00:18:21.319 --> 00:18:25.609
to hold accountable, And you have tried everything
00:18:25.609 --> 00:18:29.710
to identify who that guy is, and you're having
00:18:29.710 --> 00:18:32.630
no luck in finding the identity of that person.
00:18:33.069 --> 00:18:36.549
We have a very powerful category of tools just
00:18:36.549 --> 00:18:39.710
to get you started on that front, called facial
00:18:39.710 --> 00:18:42.470
recognition tools. So these are very powerful,
00:18:42.549 --> 00:18:44.910
but at the same time, these are very shitty tools.
00:18:45.089 --> 00:18:47.509
So you need to be careful what you upload in
00:18:47.509 --> 00:18:51.170
the process. And from here, for example, if...
00:18:51.779 --> 00:18:54.779
The reverse image searches are literally useless
00:18:54.779 --> 00:18:57.339
when you're trying to identify a person in a
00:18:57.339 --> 00:19:00.599
video or an image. So for example, if you run
00:19:00.599 --> 00:19:03.519
a search on PIMIs and it lands you on a page
00:19:03.519 --> 00:19:07.539
where you can like find more contextual information
00:19:07.539 --> 00:19:10.319
on the person like their username, email address,
00:19:10.700 --> 00:19:13.099
then from this category, you can sort of pivot
00:19:13.099 --> 00:19:16.299
or move to the social media category and people
00:19:16.299 --> 00:19:20.599
and continue your research from there. And I
00:19:20.599 --> 00:19:25.700
think One of the most important parts, and I
00:19:25.700 --> 00:19:29.359
think this kind of shows how this toolkit is
00:19:29.359 --> 00:19:32.539
evolving. And it also like highlights the roles
00:19:32.539 --> 00:19:36.319
we guardians and maintainers play is to identify
00:19:36.319 --> 00:19:39.619
and understand the challenges researchers are
00:19:39.619 --> 00:19:42.640
facing nowadays. And in this case, it would be
00:19:42.640 --> 00:19:45.519
the AI generated disinformation out there that
00:19:45.519 --> 00:19:49.480
has distorted the information landscape. So recently
00:19:49.480 --> 00:19:53.660
we have published tools on AI DeepFake detection,
00:19:54.700 --> 00:19:57.660
like DeepFake or METER, and I don't know, it's
00:19:57.660 --> 00:19:59.900
called HAYA or HAYA DeepFake voice detector,
00:20:00.039 --> 00:20:04.940
which is very useful for audio deepfakes. So
00:20:04.940 --> 00:20:10.180
if I can give an example, I use this tool. Like
00:20:10.180 --> 00:20:12.559
a year ago, there was a claim circulating that
00:20:12.559 --> 00:20:15.880
the Indian army chief admitted that the country
00:20:15.880 --> 00:20:19.440
had lost six fighter jets to Pakistan during
00:20:19.440 --> 00:20:23.779
the conflict in May 2025. And it was kind of
00:20:23.779 --> 00:20:27.779
a wild claim for an Indian officer to go publicly
00:20:27.779 --> 00:20:31.200
and say like they lost the fighter jets. And
00:20:31.200 --> 00:20:33.740
when you hear it for the first time, it does
00:20:33.740 --> 00:20:37.380
sound like it instantly clicks that this is likely
00:20:37.380 --> 00:20:40.250
to be false. But then you need to go into the
00:20:40.250 --> 00:20:43.309
verification process. And this was one of the
00:20:43.309 --> 00:20:47.250
tools that I used to identify whether his voice
00:20:47.250 --> 00:20:51.089
was real. And we eventually found out that this
00:20:51.089 --> 00:20:54.289
was a deep fake. And important to keep in mind
00:20:54.289 --> 00:20:57.470
is also like when you're using these tools, use
00:20:57.470 --> 00:20:59.869
other methods also, like reach out to experts,
00:21:00.089 --> 00:21:03.730
run reverse image searches. And so this was how
00:21:03.730 --> 00:21:06.190
this particular tool helped me from this category
00:21:06.190 --> 00:21:11.269
to debunk disinformation. And one of the most,
00:21:11.450 --> 00:21:15.250
my favorite tools, for example, if you're running
00:21:15.250 --> 00:21:17.990
reverse image searches and if you have keyframes,
00:21:18.529 --> 00:21:21.529
you need to organize that data. There's a tool
00:21:21.529 --> 00:21:25.170
called Pure Ref and we don't have a guide for
00:21:25.170 --> 00:21:29.150
it yet, but this tool is very useful for organizing
00:21:29.150 --> 00:21:32.950
your image or reference images that you might
00:21:32.950 --> 00:21:36.480
be using for geolocation. So I don't know if
00:21:36.480 --> 00:21:40.859
you can see my screen. This is how the tool looks
00:21:40.859 --> 00:21:44.880
on the dashboard of the tool. So these are the
00:21:44.880 --> 00:21:49.599
key frames from a Bellingcat video. And I can
00:21:49.599 --> 00:21:52.920
sort of arrange the images in any way I want.
00:21:53.220 --> 00:21:58.640
I could enlarge it, zoom out and rotate it. So
00:21:58.640 --> 00:22:01.259
playing around with it the way I would want.
00:22:02.470 --> 00:22:04.910
It's one of the very simplest tools that you
00:22:04.910 --> 00:22:09.089
can have in your tool kit to organize the images.
00:22:10.430 --> 00:22:12.589
But still, if you feel like this is not useful,
00:22:12.910 --> 00:22:16.730
I hope it is, because it's one of my favorite
00:22:16.730 --> 00:22:19.390
tools. It acts as a board where you can pin your
00:22:19.390 --> 00:22:24.250
images. We have something more fancy. If you
00:22:24.250 --> 00:22:27.470
prefer something fancy and sophisticated, we
00:22:27.470 --> 00:22:31.099
have... image stitching tools to create these
00:22:31.099 --> 00:22:33.640
beautiful panoramas, just in case if you need
00:22:33.640 --> 00:22:37.720
to like duplicate an image and you need a perfect
00:22:37.720 --> 00:22:45.000
picture. And Johanna already highlighted the
00:22:45.000 --> 00:22:48.180
similar tool section and the limitations. I think
00:22:48.180 --> 00:22:52.440
it's one of the very important part of this toolkit
00:22:52.440 --> 00:22:57.599
and effort by the by the maintainers. A lot of
00:22:57.599 --> 00:23:01.319
the time goes into thoroughly testing the tool
00:23:01.319 --> 00:23:04.160
and identifying what the limitations are and
00:23:04.160 --> 00:23:07.500
then ethical considerations which help you whether
00:23:07.500 --> 00:23:10.799
you should upload an image into a facial recognition
00:23:10.799 --> 00:23:14.200
tool. And if this convinces you or all of this
00:23:14.200 --> 00:23:16.259
convinces you that this tool is not the right
00:23:16.259 --> 00:23:19.180
one, you can always check out the similar tool
00:23:19.180 --> 00:23:23.869
section for your name. So this is it all from
00:23:23.869 --> 00:23:27.430
my end. And if you have any questions, please
00:23:27.430 --> 00:23:30.049
feel free to hear them in the chat and I'll pass
00:23:30.049 --> 00:23:36.170
over to Martin now. Thank you. Thank you, Anissa.
00:23:36.589 --> 00:23:41.460
Hi everyone. Thank you for the nice introduction.
00:23:42.240 --> 00:23:46.279
So over the last three years in the toolkit team,
00:23:47.119 --> 00:23:51.539
my main area has been social media and the question
00:23:51.539 --> 00:23:56.089
I keep coming back to as most of us in the toolkit
00:23:56.089 --> 00:23:59.450
team are, is very simple. Does this tool still
00:23:59.450 --> 00:24:03.279
work? It sounds like a little bit of a modest
00:24:03.279 --> 00:24:05.920
question for such an anniversary talk, but it
00:24:05.920 --> 00:24:10.660
does cover a lot of the job. A tool can have
00:24:10.660 --> 00:24:14.599
a still working website or an impressive descriptions,
00:24:15.480 --> 00:24:18.220
while the features you actually need might have
00:24:18.220 --> 00:24:21.160
stopped working or it works, but only after another
00:24:21.160 --> 00:24:24.319
login challenge or with another application or
00:24:24.319 --> 00:24:28.450
with a payment. And helping people understand
00:24:28.450 --> 00:24:32.269
those conditions has become a big part of maintaining
00:24:32.269 --> 00:24:37.109
the toolkit. And just as a clarification, we
00:24:37.109 --> 00:24:39.730
maintain independent guides for these tools,
00:24:39.890 --> 00:24:43.869
right? So maintaining an entry does not mean
00:24:43.869 --> 00:24:46.690
we have any sort of association with the tool
00:24:46.690 --> 00:24:51.009
or its developers. In contrast, that probably
00:24:51.009 --> 00:24:54.890
means that we don't have any association with
00:24:54.890 --> 00:24:57.009
the tool or its developers. Our responsibility
00:24:57.009 --> 00:25:01.519
is solely to the people using the toolkit. So
00:25:01.519 --> 00:25:05.799
they may be investigating a network or preserving
00:25:05.799 --> 00:25:09.000
evidence before it disappears. That's especially
00:25:09.000 --> 00:25:13.740
relevant in my category. And sending them down
00:25:13.740 --> 00:25:16.519
a dead end just costs time they might not have.
00:25:17.349 --> 00:25:22.170
The biggest change I've seen over the last two
00:25:22.170 --> 00:25:24.950
years is that public social media has become
00:25:24.950 --> 00:25:28.890
hard to research systematically. Just a post
00:25:28.890 --> 00:25:31.930
being visible on your phone does not mean that
00:25:31.930 --> 00:25:35.410
you can reliably search for it or collect related
00:25:35.410 --> 00:25:39.869
posts or repeat the same investigation next month.
00:25:41.369 --> 00:25:44.450
Especially with the very large online platforms
00:25:44.450 --> 00:25:49.680
like X and Meta, researchers have face increasing
00:25:49.680 --> 00:25:54.220
barriers around both technical access and the
00:25:54.220 --> 00:25:56.960
procedures for getting it if you're eligible.
00:25:58.289 --> 00:26:01.109
And tools that depend on that access have become
00:26:01.109 --> 00:26:04.589
a lot harder to maintain or have disappeared.
00:26:05.880 --> 00:26:08.880
So some disappear, others change the purpose.
00:26:08.960 --> 00:26:12.700
For example, our OXIE entry now points people
00:26:12.700 --> 00:26:20.410
towards its successor, OXIEMeNet. So checking
00:26:20.410 --> 00:26:23.430
whether a familiar name still offers the capability
00:26:23.430 --> 00:26:28.630
you remember is part of the work. And there's
00:26:28.630 --> 00:26:32.950
another example specifically with Meta's data
00:26:32.950 --> 00:26:37.670
research tools. An example was that Meta announced
00:26:37.670 --> 00:26:40.529
fees for the university hosted environment used
00:26:40.529 --> 00:26:43.329
to research metadata and Meta's own research
00:26:43.329 --> 00:26:47.210
environment remained free, which matters. changing
00:26:47.210 --> 00:26:50.410
environments, moving projects, and understanding
00:26:50.410 --> 00:26:53.569
how the conditions change, all that takes a lot
00:26:53.569 --> 00:26:59.190
of time. And sometimes the hurdles are large
00:26:59.190 --> 00:27:03.680
and intransparent. Excess can become really impractical
00:27:03.680 --> 00:27:05.940
through the accumulation of these demands, right?
00:27:06.220 --> 00:27:08.980
So a small newsroom or an independent researcher
00:27:08.980 --> 00:27:13.839
at a university with a mediocre funding may struggle.
00:27:14.579 --> 00:27:17.500
long before they reach the actual stage of research.
00:27:18.740 --> 00:27:20.619
And this is why, this is one of the reasons I
00:27:20.619 --> 00:27:22.960
want to give the open source tools some attention.
00:27:24.240 --> 00:27:28.799
So, um, zschreimer, for example, um, collects
00:27:28.799 --> 00:27:32.420
data. That's, that's a tool that collects, uh,
00:27:32.460 --> 00:27:35.640
data about supports or supported social media
00:27:35.640 --> 00:27:39.140
posts that as you browse. So while you browse,
00:27:39.160 --> 00:27:41.769
it collects the data. You can export that data
00:27:41.769 --> 00:27:45.230
later or take it into another program for analysis,
00:27:45.549 --> 00:27:48.529
Forecat, for example. And both are open source
00:27:48.529 --> 00:27:51.029
projects that are well retained. And that gives
00:27:51.029 --> 00:27:53.329
researchers a useful way to work with what they
00:27:53.329 --> 00:27:56.410
can access, even if they don't have a researcher
00:27:56.410 --> 00:27:59.970
access. We still need to explain the limits,
00:27:59.970 --> 00:28:03.210
of course, in the toolkit. So a collection of
00:28:03.210 --> 00:28:06.029
posts loaded in your browser is not a complete
00:28:06.029 --> 00:28:11.690
picture of a platform. code available also means
00:28:11.690 --> 00:28:15.250
that people can inspect how your collection works.
00:28:15.650 --> 00:28:18.549
They can identify mistakes, they can adapt the
00:28:18.549 --> 00:28:22.430
tool to new questions or help prepare it whenever
00:28:22.430 --> 00:28:27.900
a platform inevitably changes, right? And for
00:28:27.900 --> 00:28:31.900
investigative work, being able to explain how
00:28:31.900 --> 00:28:37.440
you obtain the result really matters. Open code
00:28:37.440 --> 00:28:41.079
gives us just more opportunity to check the tools
00:28:41.079 --> 00:28:44.019
that we depend on. It also gives a project a
00:28:44.019 --> 00:28:46.660
chance to continue if the original developer
00:28:46.660 --> 00:28:49.079
or the original researcher has moved on to something
00:28:49.079 --> 00:28:52.039
else. Of course, somebody still has to do that
00:28:52.039 --> 00:28:54.799
work. An open source software needs maintenance,
00:28:55.539 --> 00:28:59.000
documentation needs funding. I've seen this in
00:28:59.000 --> 00:29:03.079
a couple of social media tools and other tools
00:29:03.079 --> 00:29:06.390
that have. that went through different rounds
00:29:06.390 --> 00:29:10.349
of funding and have changed hands and teams over
00:29:10.349 --> 00:29:13.390
the course of the last three years. And we should
00:29:13.390 --> 00:29:16.349
support the people doing it, including through
00:29:16.349 --> 00:29:19.470
useful bug reports, contributions, through credit.
00:29:20.970 --> 00:29:25.369
Of course, OpenCode cannot restore access to
00:29:25.369 --> 00:29:29.329
data that a platform has closed off. And we need
00:29:29.329 --> 00:29:35.740
both usable tools and usable access. There have
00:29:35.740 --> 00:29:38.759
been encouraging developments, too. So we added,
00:29:38.759 --> 00:29:41.880
for example, Blue Sky Tools in December 2024.
00:29:42.319 --> 00:29:46.779
We published guides the following March. Blue
00:29:46.779 --> 00:29:48.940
Sky is an open architecture that has created
00:29:48.940 --> 00:29:51.160
a lot of room for people to build new ways of
00:29:51.160 --> 00:29:53.319
exploring social connections. That's really great.
00:29:55.250 --> 00:29:58.190
Looking through the toolkit's GitHub history,
00:29:58.390 --> 00:30:05.349
you can see both sides of these two gears. documenting
00:30:05.349 --> 00:30:09.569
new possibilities and keeping up with methods
00:30:09.569 --> 00:30:14.049
that have become less reliable. That brings me
00:30:14.049 --> 00:30:18.230
a little bit to the European rules and what platform
00:30:18.230 --> 00:30:21.690
transparency means. So I really support the EU's
00:30:21.690 --> 00:30:24.789
commitment to platform transparency, but I want
00:30:24.789 --> 00:30:27.869
the rules to be clearer in practice about who
00:30:27.869 --> 00:30:32.349
qualifies, what they can request. how they can
00:30:32.349 --> 00:30:35.049
challenge a refusal. So for me, as an academic
00:30:35.049 --> 00:30:39.349
researcher, it might be really easy to get, not
00:30:39.349 --> 00:30:43.349
that it might be possible to get access, but
00:30:43.349 --> 00:30:47.170
it might be a lot harder for an independent researcher,
00:30:47.349 --> 00:30:49.900
freelancer, or a newsroom. There are different
00:30:49.900 --> 00:30:51.880
routes for public data and for better access
00:30:51.880 --> 00:30:55.000
to non -public data with different requirements,
00:30:55.259 --> 00:30:57.400
like for journalists, freelancers, for civil
00:30:57.400 --> 00:31:00.170
society researchers as well. Finding the route
00:31:00.170 --> 00:31:02.470
that actually applies to them should be much
00:31:02.470 --> 00:31:08.470
easier. A platform should not be able to make
00:31:08.470 --> 00:31:13.309
a right to research unusable to an exhausting
00:31:13.309 --> 00:31:15.430
application process. And I've been struggling
00:31:15.430 --> 00:31:17.990
with this in a couple of my toolkit entries a
00:31:17.990 --> 00:31:21.990
lot, keeping up with the changes in those application
00:31:21.990 --> 00:31:24.970
processes. And enforcement also matters here.
00:31:25.170 --> 00:31:29.680
In October 2025, for example, a commission Preliminary
00:31:29.680 --> 00:31:33.700
found that Metta and TikTok were failing to provide
00:31:33.700 --> 00:31:36.640
adequate research access. In December, it find
00:31:36.640 --> 00:31:40.240
X for transparency breaches that included restricting
00:31:40.240 --> 00:31:43.380
access to public data for researchers. And of
00:31:43.380 --> 00:31:45.400
course, those are welcome steps and important
00:31:45.400 --> 00:31:49.279
steps. And we need consistent follow through.
00:31:49.740 --> 00:31:52.500
Researchers need ways to resolve access problems
00:31:52.500 --> 00:31:54.920
while their investigations are still timely.
00:31:57.700 --> 00:32:03.140
The news cycle is short. The measure of success
00:32:03.140 --> 00:32:07.799
in general, but also for us at the toolkit is
00:32:07.799 --> 00:32:11.059
whether people can actually do the research on
00:32:11.059 --> 00:32:13.400
a legislative level, but also practically for
00:32:13.400 --> 00:32:16.740
us as a toolkit team. And that's also why, for
00:32:16.740 --> 00:32:18.779
example, the transatlantic argument concerns
00:32:18.779 --> 00:32:21.599
below, but the US administration has challenged
00:32:21.599 --> 00:32:25.640
European rules as unfair treatment of American
00:32:25.640 --> 00:32:28.069
companies. So what it's actually usually about
00:32:28.069 --> 00:32:31.720
why these changes happen is money. I want Europe
00:32:31.720 --> 00:32:35.599
to defend independent scrutiny while making its
00:32:35.599 --> 00:32:38.279
own rules clear and workable. Researchers on
00:32:38.279 --> 00:32:44.000
both sides will benefit when they can examine
00:32:44.000 --> 00:32:48.819
powerful platforms without depending on the platform's
00:32:48.819 --> 00:32:50.740
goodwill. Because if I've seen anything in the
00:32:50.740 --> 00:32:54.000
last two years is that a platform's goodwill
00:32:54.000 --> 00:32:59.180
is something that we cannot depend on. So for
00:32:59.180 --> 00:33:02.160
me, the toolkit is a practical contribution to
00:33:02.160 --> 00:33:07.359
that effort. Every guide we review, we test,
00:33:07.380 --> 00:33:11.700
we write. Every limitation we explain and also
00:33:11.700 --> 00:33:15.869
broken methods that we flag do help someone else
00:33:15.869 --> 00:33:19.690
investigate. Every open source project that we
00:33:19.690 --> 00:33:22.890
help people understand gives that work a better
00:33:22.890 --> 00:33:26.069
chance of continuing. or gives the newsroom a
00:33:26.069 --> 00:33:29.710
possibility that they might not have had otherwise
00:33:29.710 --> 00:33:33.109
because they work out of a country that isn't
00:33:33.880 --> 00:33:36.200
equipped in the same way as say the New York
00:33:36.200 --> 00:33:43.579
Times or the New York Post. To finish it up,
00:33:43.740 --> 00:33:46.140
that's what I've valued over those two years
00:33:46.140 --> 00:33:49.440
and on the team, like sharing what we learn and
00:33:49.440 --> 00:33:53.180
helping to keep all this work possible. That's
00:33:53.180 --> 00:33:57.509
it for me. Thank you so much. Thank you. Thanks
00:33:57.509 --> 00:34:00.130
first, Johanna, going through what the toolkit
00:34:00.130 --> 00:34:04.369
means, and then both of the guardians explaining
00:34:04.369 --> 00:34:07.269
what makes their categories special and also
00:34:07.269 --> 00:34:10.570
those marks of success. We're going to go into
00:34:10.570 --> 00:34:13.269
the Q &A now. We've already got a few questions
00:34:13.269 --> 00:34:15.429
in the chat, but I wanted to start off with this
00:34:15.429 --> 00:34:18.889
question. Perhaps there's people listening, for
00:34:18.889 --> 00:34:21.869
example, who perhaps are developing their own
00:34:21.869 --> 00:34:26.750
tools. or are aware of tools that they find useful.
00:34:27.349 --> 00:34:31.130
There's now an enormous amount of OSINT tools
00:34:31.130 --> 00:34:33.909
out there and the toolkit says that inclusion
00:34:33.909 --> 00:34:36.550
doesn't necessarily constitute an endorsement
00:34:36.550 --> 00:34:39.030
and that some tools have significant caveats.
00:34:39.349 --> 00:34:42.530
So what makes a tool worthy of inclusion in the
00:34:42.530 --> 00:34:44.769
toolkit and how do you evaluate whether a tool
00:34:44.769 --> 00:34:47.130
is suitable for inclusion in the first place?
00:34:48.090 --> 00:34:52.449
I'm happy to start talking about that. So this
00:34:52.449 --> 00:34:57.170
is a very long and ongoing discussion we have
00:34:57.170 --> 00:35:00.570
in the project because on the one hand, we often
00:35:00.570 --> 00:35:03.489
want to include everything, but it's just not
00:35:03.489 --> 00:35:07.289
possible. So we always discuss, okay, should
00:35:07.289 --> 00:35:12.110
we change our criteria, what we accept and what
00:35:12.110 --> 00:35:14.630
we don't? So I can tell you how it all started.
00:35:14.710 --> 00:35:17.800
Since this is a volunteer -driven project, we
00:35:17.800 --> 00:35:20.840
really started by just asking every volunteer
00:35:20.840 --> 00:35:23.559
what they are most interested in. So the tools
00:35:23.559 --> 00:35:27.179
you see in the toolkit are mostly coming from
00:35:27.179 --> 00:35:30.719
suggestions from the people who also test these
00:35:30.719 --> 00:35:34.219
tools and had a motivation or an interest to
00:35:34.219 --> 00:35:37.300
test these tools. But at the moment, we are discussing
00:35:37.300 --> 00:35:42.179
a lot in the Bellingcat team, those people of
00:35:42.179 --> 00:35:45.719
us who spend time within the toolkit project.
00:35:45.989 --> 00:35:48.690
and we might want to change that in the long
00:35:48.690 --> 00:35:55.570
run and to find a way to let the broader communities
00:35:55.570 --> 00:35:59.630
we have suggest us tools in a more structured
00:35:59.630 --> 00:36:02.429
way. We are still figuring out how to do that
00:36:02.429 --> 00:36:05.730
exactly, so it's not yet happening, but we want
00:36:05.730 --> 00:36:09.510
in the future that we get signals from the wider
00:36:09.510 --> 00:36:12.949
open source researcher community what type of
00:36:12.949 --> 00:36:17.199
tools they would like to see in the toolkit most
00:36:17.199 --> 00:36:20.900
because we receive a lot of emails and messages
00:36:20.900 --> 00:36:23.519
from individuals who want really a specific tool
00:36:23.519 --> 00:36:26.880
to be included. But you know that would just
00:36:26.880 --> 00:36:29.320
be the opinion of one person. So we are working
00:36:29.320 --> 00:36:33.400
on a potential system to get the suggestions
00:36:33.400 --> 00:36:38.119
from more people in and hopefully you will see
00:36:38.119 --> 00:36:42.440
the system come into place. Maybe not this year,
00:36:42.440 --> 00:36:45.429
but hopefully next year. And as I wrote in the
00:36:45.429 --> 00:36:48.230
chat, there is a space, for example, in the Discord
00:36:48.230 --> 00:36:51.849
server, if you do want to share tools, we have
00:36:51.849 --> 00:36:55.329
a space for established tools that have been
00:36:55.329 --> 00:36:58.750
peer reviewed called hashtag tools and sites.
00:36:58.909 --> 00:37:01.409
And then there's also a space if you are developing
00:37:01.409 --> 00:37:04.190
your own tool and you're looking for advice,
00:37:04.630 --> 00:37:09.659
reviews, feedback, that is hashtag. tool developments,
00:37:09.739 --> 00:37:12.880
which is a really growing channel actually. We've
00:37:12.880 --> 00:37:15.280
got quite a few people in here working on their
00:37:15.280 --> 00:37:18.659
own tools, which is really, really cool. FD in
00:37:18.659 --> 00:37:21.300
the chat says, I used to work as an electrician
00:37:21.300 --> 00:37:24.059
and lots of people, myself included, become over
00:37:24.059 --> 00:37:27.880
-focused on tools and neglect skills. I'm new
00:37:27.880 --> 00:37:30.199
to investigating, but I found myself falling
00:37:30.199 --> 00:37:33.440
into that mindset again. Does anyone have advice
00:37:33.440 --> 00:37:37.739
on getting the balance right? Do any of our resident
00:37:37.739 --> 00:37:39.920
investigators here, I'm talking to you Anita
00:37:39.920 --> 00:37:42.039
and Martin, want to answer that one? How do you
00:37:42.039 --> 00:37:46.210
balance the need for skills versus tools? For
00:37:46.210 --> 00:37:49.409
me, it really helps to have a project. You need
00:37:49.409 --> 00:37:51.550
something that you want to use the tools on.
00:37:53.719 --> 00:37:58.880
And I don't think it's wrong to try as many tools
00:37:58.880 --> 00:38:02.119
as possible. I'm a tool person myself, but really
00:38:02.119 --> 00:38:05.019
learning appears when you apply your knowledge
00:38:05.019 --> 00:38:10.480
to something. So that would be my advice is get
00:38:10.480 --> 00:38:14.360
a project to use those tools and then try it.
00:38:16.460 --> 00:38:18.980
I don't know if you want to add something Anisa,
00:38:20.139 --> 00:38:23.219
but otherwise I do and please add to that. I
00:38:23.219 --> 00:38:27.219
fully agree to what you said, Martin. What I
00:38:27.219 --> 00:38:31.239
often see is that people are nearly scared of
00:38:31.239 --> 00:38:34.460
tools when they are beginning to do open source
00:38:34.460 --> 00:38:38.239
research. I reach out and say, look, please,
00:38:38.460 --> 00:38:40.940
which tools should I learn first? Give me a list
00:38:40.940 --> 00:38:45.159
and I learn them all. I would really try to reverse
00:38:45.159 --> 00:38:48.679
that approach. Very similar to what Martin said,
00:38:48.679 --> 00:38:51.639
because I feel if you have a clear in your working
00:38:51.639 --> 00:38:56.539
on and on the way you learn tools, it can take
00:38:56.539 --> 00:38:59.179
out the fear for some people. If you're a very
00:38:59.179 --> 00:39:01.639
structured person and you want to go through
00:39:01.639 --> 00:39:04.420
tools one after the other, that is also fine.
00:39:04.699 --> 00:39:08.980
But I specifically recommend this approach to
00:39:08.980 --> 00:39:12.340
attach it to a research case to those who get
00:39:12.340 --> 00:39:17.699
super fearful by the sheer amount of tools that
00:39:17.699 --> 00:39:20.159
are out there because it can be super intimidating
00:39:20.159 --> 00:39:23.179
and I think it doesn't have to be. You don't
00:39:23.179 --> 00:39:26.699
need to know all the tools. I would say nobody
00:39:26.699 --> 00:39:29.599
in the project team and the toolkit project team
00:39:29.599 --> 00:39:33.030
here knows all the tools in the toolkit, right?
00:39:33.269 --> 00:39:36.030
We are constantly all learning from each other
00:39:36.030 --> 00:39:39.469
and that is fine. And that's also the strength
00:39:39.469 --> 00:39:44.570
of it. So let first of all, let the thought go
00:39:44.570 --> 00:39:47.329
that you need to know all of them. That would
00:39:47.329 --> 00:39:53.289
be my suggestion. Yeah, I think to add to what
00:39:53.289 --> 00:39:57.150
Martin and Johanna said, it's important to keep
00:39:57.150 --> 00:39:59.989
the tools and skills separate. I often get a
00:39:59.989 --> 00:40:01.750
question like what are the tools that can get
00:40:01.750 --> 00:40:05.610
a geolocation done when in fact skills are separate
00:40:05.610 --> 00:40:08.150
from the tools and what works best is to apply
00:40:08.150 --> 00:40:11.190
them in your research problems and over time
00:40:11.190 --> 00:40:13.530
you will get the hang of using the right tool
00:40:13.530 --> 00:40:16.969
relevant to your research. Amazing tips there.
00:40:17.449 --> 00:40:20.449
I also liked Martin's in the comments. Also keep
00:40:20.449 --> 00:40:22.510
in mind that your most important tool is probably
00:40:22.510 --> 00:40:26.699
or always a spreadsheet that is very true. Although
00:40:26.699 --> 00:40:29.179
our mod Subtleknife disagrees with you and says
00:40:29.179 --> 00:40:35.739
number one is your brain power. Amazing. I wanted
00:40:35.739 --> 00:40:39.320
to ask, where do you think the biggest blind
00:40:39.320 --> 00:40:43.099
spots currently are in OSINT in your respective
00:40:43.099 --> 00:40:46.639
category topics? So social media and image and
00:40:46.639 --> 00:40:49.659
video. Are there things researchers are routinely
00:40:50.219 --> 00:40:52.579
Assuming that they can find actually require
00:40:52.579 --> 00:40:55.420
much more caution in your two categories. Yeah,
00:40:55.519 --> 00:40:57.420
where are the biggest blind spots that you see
00:40:57.420 --> 00:41:01.610
when reviewing these tools? In my category, I
00:41:01.610 --> 00:41:03.889
haven't come across and you would need to think
00:41:03.889 --> 00:41:07.030
it out. Maybe Martin has something to add to
00:41:07.030 --> 00:41:10.599
that, but it's a good question. That's a good
00:41:10.599 --> 00:41:15.760
question, especially with social media. It's
00:41:15.760 --> 00:41:19.079
because even when you use a search function,
00:41:19.119 --> 00:41:25.869
even when you use some kind of tool, you are
00:41:25.869 --> 00:41:30.050
always being shown what the search function delivers.
00:41:30.670 --> 00:41:33.469
You never get a full picture. It doesn't matter
00:41:33.469 --> 00:41:40.039
if you have what kind of access you use. uh,
00:41:41.909 --> 00:41:44.570
your processing of, of the event, your processing
00:41:44.570 --> 00:41:47.230
of what, what you read, what you see that leads
00:41:47.230 --> 00:41:50.550
you to the things that you discover. And that's
00:41:50.550 --> 00:41:52.309
something that you would be always have to keep
00:41:52.309 --> 00:41:56.409
in mind that you're only seeing, um, and I'm
00:41:56.409 --> 00:41:58.989
not saying that this is some sort of conspiracy
00:41:58.989 --> 00:42:00.969
that they only, that you only see what you want
00:42:00.969 --> 00:42:03.630
to see, but you have to, you have to really be
00:42:03.630 --> 00:42:07.789
aware that, um, there is a, there, there is a
00:42:07.789 --> 00:42:10.920
some, that what you see doesn't, isn't that necessarily
00:42:10.920 --> 00:42:16.449
a true reflection of what's out there. Maybe
00:42:16.449 --> 00:42:21.730
I can add, I feel for video, the challenge currently
00:42:21.730 --> 00:42:25.710
are all these deepfake detection tools that are
00:42:25.710 --> 00:42:28.510
coming out and we also start including them into
00:42:28.510 --> 00:42:30.829
toolkit and for a reason we want to include them
00:42:30.829 --> 00:42:34.110
of course. But often the challenge is that they
00:42:34.110 --> 00:42:37.949
often give us a percentage probability of how
00:42:37.949 --> 00:42:42.309
likely it is that a video is a deep fake. And
00:42:42.309 --> 00:42:47.269
that is not super useful for researchers. What
00:42:47.269 --> 00:42:50.710
we really need is an explanation why something
00:42:50.710 --> 00:42:54.210
is a deep fake or not. And many tools are not
00:42:54.210 --> 00:42:58.190
focused on that yet. So I feel that creates a
00:42:58.190 --> 00:43:02.610
lot of confusion at the moment. I don't say they
00:43:02.610 --> 00:43:05.809
are not useful, use them, but be aware that you
00:43:05.809 --> 00:43:11.269
also need other ways of verification. I feel
00:43:11.269 --> 00:43:14.670
we all are closely observing how these tools
00:43:14.670 --> 00:43:17.989
will be changing and hopefully more tools will
00:43:17.989 --> 00:43:20.889
be focusing on not giving us just a probability
00:43:20.889 --> 00:43:24.269
score, but a reasoning of how they came to the
00:43:24.269 --> 00:43:30.719
conclusion and why. Yeah, it's important when
00:43:30.719 --> 00:43:34.539
you're using deep fake detection tools to find
00:43:34.539 --> 00:43:37.039
other ways of verifying our deep fake media,
00:43:37.300 --> 00:43:41.340
like reach out to experts for analysis. If you're
00:43:41.340 --> 00:43:45.960
using one tool, add two or more, three or more
00:43:45.960 --> 00:43:49.019
to your toolkit for analysis to double check
00:43:49.019 --> 00:43:53.139
the results and still look for visual clues in
00:43:53.139 --> 00:43:55.760
the video, like AI and these artifacts in the
00:43:55.760 --> 00:43:59.199
video. So that can also be helpful. And at the
00:43:59.199 --> 00:44:02.170
same time, look for any context information you
00:44:02.170 --> 00:44:06.130
can find from news platforms, from fact checking
00:44:06.130 --> 00:44:09.489
organizations who are working on these claims
00:44:09.489 --> 00:44:12.389
and videos. We've had another question. What
00:44:12.389 --> 00:44:16.050
has been the most common reason for removing
00:44:16.050 --> 00:44:20.929
a tool? Broken functionality or a paywall? Has
00:44:20.929 --> 00:44:23.610
there been kind of common reasons tools have
00:44:23.610 --> 00:44:27.849
been removed from the toolkit? Oh, yeah, that's
00:44:27.849 --> 00:44:31.030
a question for me because I typically have to
00:44:31.030 --> 00:44:34.550
remove them and it's always a difficult decision.
00:44:35.030 --> 00:44:38.489
So it's not such a difficult decision if the
00:44:38.489 --> 00:44:41.510
tool just stops functioning. And honestly, the
00:44:41.510 --> 00:44:46.730
top reason that the often are a bit attached
00:44:46.730 --> 00:44:48.869
to the tools that we have listed in the toolkit.
00:44:49.170 --> 00:44:53.150
And we already see for some time that it's getting
00:44:53.150 --> 00:44:56.650
less useful, something stopped working, but then
00:44:56.650 --> 00:44:59.409
it's like, oh, well, two things are still working.
00:44:59.690 --> 00:45:02.210
While five don't work anymore, we still keep
00:45:02.210 --> 00:45:04.949
it. And then at some point, there is a point
00:45:04.949 --> 00:45:09.170
where I hopefully remove it, but I feel I often
00:45:09.170 --> 00:45:12.449
remove. them a bit too late, but that's really
00:45:12.449 --> 00:45:19.210
the main reason for removing a tool. Other reasons
00:45:19.210 --> 00:45:22.739
I feel sometimes, since we are still debating
00:45:22.739 --> 00:45:26.960
internally what type of tools we include, I mentioned
00:45:26.960 --> 00:45:30.679
earlier that it is mainly based on people's interests
00:45:30.679 --> 00:45:33.800
so far, but sometimes we then end up having a
00:45:33.800 --> 00:45:37.920
few tools that are maybe a bit too niche and
00:45:37.920 --> 00:45:42.000
there might be like five other very similar niche
00:45:42.000 --> 00:45:45.000
tours out there. So either at some point we would
00:45:45.000 --> 00:45:48.079
need to include these five others, super niche.
00:45:48.400 --> 00:45:51.659
And I mean, we can be niche, but I mean really
00:45:51.659 --> 00:45:56.500
niche. We need to add five others too, or at
00:45:56.500 --> 00:45:59.579
some point maybe we take this one out. So these
00:45:59.579 --> 00:46:05.860
are, I would say, the most common common things
00:46:05.860 --> 00:46:09.079
and sometimes a tool name also just changes,
00:46:09.159 --> 00:46:13.219
but it's listed under specific name. Then we
00:46:13.219 --> 00:46:15.559
ask ourselves, okay, everyone knows the tool
00:46:15.559 --> 00:46:18.539
under the old name. When do we switch to the
00:46:18.539 --> 00:46:21.280
new one? Because people won't find it for a while.
00:46:21.280 --> 00:46:24.260
So we keep the old name for a while. So yeah,
00:46:24.260 --> 00:46:26.400
I don't want to bore you with all the details,
00:46:26.420 --> 00:46:30.219
but these changes I would say are the most common
00:46:30.219 --> 00:46:33.679
reasons to take a tool out. Maybe sometimes we
00:46:33.679 --> 00:46:36.940
find that a tool is shady, that can also happen
00:46:36.940 --> 00:46:40.159
that a research comes out and then we are debating
00:46:40.159 --> 00:46:43.380
internally about that. But I think Martin wants
00:46:43.380 --> 00:46:45.880
to add maybe on this point. Go ahead, Martin.
00:46:46.059 --> 00:46:49.500
Yeah, exactly. I just wanted to say that because
00:46:49.500 --> 00:46:53.719
we do have to walk a little bit of a path when
00:46:53.719 --> 00:46:56.179
selecting the toolkit. So because a lot of the
00:46:56.179 --> 00:47:00.679
tools that we use and maintain walk on the fine
00:47:00.679 --> 00:47:05.239
line of the terms of service, which we Sometimes
00:47:05.239 --> 00:47:10.079
I personally see often as more suggestions than
00:47:10.079 --> 00:47:14.019
legal binding contracts, but there are some tools
00:47:14.019 --> 00:47:18.000
that we have. I can think of at least one that
00:47:18.000 --> 00:47:25.119
I've personally said we need to remove that used
00:47:25.119 --> 00:47:28.960
a fraudulent payment tech text. It was definitely
00:47:28.960 --> 00:47:31.579
way over the line in terms of terms of service,
00:47:31.760 --> 00:47:35.480
but that also used fraudulent payment methods.
00:47:36.039 --> 00:47:38.659
And so this is generally a line that we need
00:47:38.659 --> 00:47:44.150
to walk. How invasive are those tools? And in
00:47:44.150 --> 00:47:46.590
the end, for me, it's personally a question of
00:47:46.590 --> 00:47:51.730
if I do an investigation and I need to tell people
00:47:51.730 --> 00:47:55.610
what tool I used, is this a tool that I would
00:47:55.610 --> 00:48:01.369
be happy to be associated with? And that's my
00:48:01.369 --> 00:48:03.409
personal line, but of course, it's always a discussion.
00:48:05.289 --> 00:48:09.309
And this is why regular checkbacks on the tools
00:48:09.309 --> 00:48:12.210
that you've got in your categories are important,
00:48:12.449 --> 00:48:16.510
right? How do you keep updated with those updates
00:48:16.510 --> 00:48:18.710
to tools that you perhaps are happy to have in
00:48:18.710 --> 00:48:21.409
the toolkit but may have changed slightly? I'm
00:48:21.409 --> 00:48:23.630
thinking particularly of the social media tools
00:48:23.630 --> 00:48:26.210
right now. How do you keep on top of that? And
00:48:26.210 --> 00:48:30.159
if somebody who's listening perhaps notices one
00:48:30.159 --> 00:48:32.840
of the tools listed in the toolkit has changed,
00:48:33.159 --> 00:48:35.380
broken, disappeared, or introduced any limitation.
00:48:36.079 --> 00:48:38.699
What's the best way for people in the community
00:48:38.699 --> 00:48:42.739
to flag that to you guys as toolkit guardians?
00:48:43.320 --> 00:48:46.219
Maybe I'll just start with the second question.
00:48:46.380 --> 00:48:50.579
If you notice that a tool stops working or something
00:48:50.579 --> 00:48:53.360
else is up with the tool, you can always write
00:48:53.360 --> 00:48:57.320
an email to toolkit at bellingcat .com. And we
00:48:57.320 --> 00:48:59.699
react to that, right? If you let us know, look,
00:48:59.820 --> 00:49:02.960
I tested this tool. It's really not working anymore.
00:49:03.000 --> 00:49:05.579
And it happens. And we are super grateful in
00:49:05.579 --> 00:49:08.699
case some of you who listen already did that.
00:49:08.820 --> 00:49:11.820
It really helps us a lot. We actually remove
00:49:11.820 --> 00:49:15.119
the tool or adapt information about something
00:49:15.119 --> 00:49:19.360
not working. And for how to best update it, I
00:49:19.360 --> 00:49:23.420
hand over. Maybe, Anita, with the video category,
00:49:23.940 --> 00:49:27.809
a lot is also changing, right? I think both of
00:49:27.809 --> 00:49:30.889
you are the best to answer that question. In
00:49:30.889 --> 00:49:35.849
terms of how we stay updated, we update the tools
00:49:35.849 --> 00:49:39.150
bi -monthly and for that we keep checking the
00:49:39.150 --> 00:49:43.010
tools when the deadline approaches and we test
00:49:43.010 --> 00:49:45.750
them regularly in that sense. We visit their
00:49:45.750 --> 00:49:49.409
websites if the tool provider has an update about
00:49:49.409 --> 00:49:53.380
their tool. And most of the tools don't change
00:49:53.380 --> 00:49:56.800
that much, like they stay the same. Yes, there
00:49:56.800 --> 00:49:59.960
have been times when a certain tool stops working
00:49:59.960 --> 00:50:03.219
and we have removed those from the toolkit. But
00:50:03.219 --> 00:50:07.440
yeah, testing them out, like following investigative
00:50:07.440 --> 00:50:10.039
researchers out there on X and other platforms,
00:50:10.219 --> 00:50:13.820
if they notice any change, it's a good way of
00:50:13.820 --> 00:50:17.400
like keeping yourself updated on that print.
00:50:17.769 --> 00:50:23.889
And another aspect is maybe also that we often
00:50:23.889 --> 00:50:30.650
try to really look if important parts of the
00:50:30.650 --> 00:50:33.150
tools change. So for example, what I always do
00:50:33.150 --> 00:50:36.269
is for command line tools or for open source
00:50:36.269 --> 00:50:38.869
tools, I keep a close eye on the issue tracker.
00:50:39.510 --> 00:50:41.949
So when a tool stops working, obviously people
00:50:41.949 --> 00:50:45.190
will complain at GitHub. That's always important
00:50:45.190 --> 00:50:48.730
information for me. I always, on the tools that
00:50:48.730 --> 00:50:52.010
I maintain, I keep a check on if anything changes
00:50:52.010 --> 00:50:56.070
in the privacy policy, for example, or in the
00:50:56.070 --> 00:50:58.909
fine print, basically. So whenever the fine print
00:50:58.909 --> 00:51:02.929
changes, I get an alert. The cookie policy, whatever.
00:51:04.550 --> 00:51:08.030
And the other thing is, of course, there is news.
00:51:08.250 --> 00:51:13.239
You can monitor whether a tool comes up in the
00:51:13.239 --> 00:51:15.760
Google search results. So there's Google Alerts.
00:51:16.400 --> 00:51:21.820
But yeah, we do not only try to keep up in terms
00:51:21.820 --> 00:51:24.940
of trying it. Does the tool still work? But also
00:51:24.940 --> 00:51:29.030
in terms of, is there maybe something that we
00:51:29.030 --> 00:51:33.139
miss if we just look at the tool itself? Thank
00:51:33.139 --> 00:51:35.780
you both. We've only got five minutes left, so
00:51:35.780 --> 00:51:38.199
we've only got time for a couple more questions,
00:51:38.360 --> 00:51:41.210
one or two. Nick asked, do you think about the
00:51:41.210 --> 00:51:44.210
toolkit as generating visible transparency and
00:51:44.210 --> 00:51:46.389
trust in Bellingcat as a broader organization?
00:51:46.590 --> 00:51:49.730
Perhaps this is one for you, Johanna? I'm always
00:51:49.730 --> 00:51:51.909
blown away by how transparent you guys always
00:51:51.909 --> 00:51:55.019
manage to be. It's incredible. Very nice words,
00:51:55.280 --> 00:51:58.260
Nick. Thank you. Thank you. Yeah, thank you for
00:51:58.260 --> 00:52:00.400
this question. At least that's the goal, right?
00:52:02.070 --> 00:52:05.130
I really want that everyone can do this type
00:52:05.130 --> 00:52:10.250
of research and not only people working at organizations
00:52:10.250 --> 00:52:15.469
with budgets for tools and maybe the funding
00:52:15.469 --> 00:52:19.849
to attend courses, right? It's great if you have
00:52:19.849 --> 00:52:23.429
all these opportunities, but open source research
00:52:23.429 --> 00:52:26.449
really gives us all the chance to participate
00:52:26.449 --> 00:52:30.429
and we want to make this possible. And this is
00:52:30.429 --> 00:52:34.159
also why I invest my time in the toolkit and
00:52:34.159 --> 00:52:37.179
it takes a lot of time for all of us here to
00:52:37.179 --> 00:52:41.900
maintain this resource. And as some of you maybe
00:52:41.900 --> 00:52:45.539
know, maintaining some times it can be a very
00:52:45.539 --> 00:52:48.679
tedious task in the background that nobody sees
00:52:48.679 --> 00:52:51.840
and people can take for granted. So I want to
00:52:51.840 --> 00:52:54.500
thank one more time all the volunteers contributing
00:52:54.500 --> 00:52:57.500
to this project for two years, because it's not
00:52:57.500 --> 00:53:00.380
always as visible to anyone what type of effort
00:53:00.380 --> 00:53:03.739
goes into it. But the goal is really that we
00:53:03.739 --> 00:53:07.099
want that everyone can do open source investigations.
00:53:07.380 --> 00:53:09.860
And of course, tools are only a part of it, but
00:53:09.860 --> 00:53:13.739
they are still an important part of it. And if
00:53:13.739 --> 00:53:16.719
we help people understand how they can use tools
00:53:16.719 --> 00:53:20.639
and what type of investigations or specific tool
00:53:20.639 --> 00:53:24.159
makes sense and for what other type of investigation
00:53:24.159 --> 00:53:29.289
may be less though, I feel we can Make a positive
00:53:29.289 --> 00:53:31.909
impact on the field, right? And it's not important
00:53:31.909 --> 00:53:34.230
whether Bellingcat does the investigation based
00:53:34.230 --> 00:53:37.510
on these tools or someone else. We really want
00:53:37.510 --> 00:53:43.090
that important investigations come out and also
00:53:43.090 --> 00:53:47.230
based on tools and that people know how to use
00:53:47.230 --> 00:53:51.210
these tools. So this is really our daily motivation.
00:53:52.110 --> 00:53:54.710
And yeah, we try to be as transparent as possible
00:53:54.710 --> 00:53:58.559
because It's not about Bellingcat, right? It's
00:53:58.559 --> 00:54:02.380
not about us as a team. It's about opening up
00:54:02.380 --> 00:54:05.820
the space that was only available to professional
00:54:05.820 --> 00:54:08.460
journalists in the past and other researchers
00:54:08.460 --> 00:54:11.320
with the long training, everyone who can teach
00:54:11.320 --> 00:54:14.820
this themselves. This is why we are all here.
00:54:15.079 --> 00:54:18.440
And I hope the toolkit helps with that. It's
00:54:18.440 --> 00:54:22.179
the open source mindset, isn't it? And as Craig
00:54:22.179 --> 00:54:25.610
Silverman mentioned last week in his talk, He
00:54:25.610 --> 00:54:29.550
mentioned that the indicators toolkit search
00:54:29.550 --> 00:54:32.789
also includes the Bellingcat tools and the Bellingcat
00:54:32.789 --> 00:54:36.190
toolkit, which is amazing. Shows kind of how
00:54:36.190 --> 00:54:38.449
we are kind of willing to share the resources
00:54:38.449 --> 00:54:40.849
that we've collected with other organizations
00:54:40.849 --> 00:54:45.010
as well. And Craig gave that a shout out in last
00:54:45.010 --> 00:54:47.710
week's talk, which you can listen back to on
00:54:47.710 --> 00:54:50.210
our RSS feed or by searching Stage Talks with
00:54:50.210 --> 00:54:52.849
Bellingcat. Okay, final, final question because
00:54:52.849 --> 00:54:56.670
we are almost at time. and try to answer this
00:54:56.670 --> 00:54:59.570
quickly if possible. But I'd love to hear an
00:54:59.570 --> 00:55:02.510
answer from all three of you. If money was no
00:55:02.510 --> 00:55:05.730
object, if we suddenly got investment for the
00:55:05.730 --> 00:55:07.809
toolkit, a large investment for the toolkit,
00:55:08.269 --> 00:55:10.889
what feature would you want to add to the toolkit
00:55:10.889 --> 00:55:13.329
that doesn't currently exist? It could be a feature
00:55:13.329 --> 00:55:16.190
specifically for your category, or it could be
00:55:16.190 --> 00:55:20.110
a feature for the general toolkit itself. Johanna,
00:55:20.110 --> 00:55:23.110
do you want to go first? And then we can switch
00:55:23.110 --> 00:55:25.809
to the other two. This is such a good question.
00:55:26.369 --> 00:55:29.670
For me, it's a very simple answer, but I think
00:55:29.670 --> 00:55:33.570
it would have a big impact. I would like to work
00:55:33.570 --> 00:55:37.909
with way more volunteers, which we could group
00:55:37.909 --> 00:55:42.750
according to specific tool interests and just
00:55:42.750 --> 00:55:46.469
spend way more time with them to compare different
00:55:46.469 --> 00:55:48.889
tools that are similar. Because when we compare
00:55:48.889 --> 00:55:52.989
tools, we find out so many things. To which extent
00:55:52.989 --> 00:55:55.889
a tool is shady, what works, what doesn't. If
00:55:55.889 --> 00:55:58.150
you compare, you find out more. So that would
00:55:58.150 --> 00:56:01.630
be my dream, way more comparing. And it requires
00:56:01.630 --> 00:56:05.449
a lot of resources and a lot of effort. Yes,
00:56:05.469 --> 00:56:08.269
I'm going to jump in because I think that's such
00:56:08.269 --> 00:56:10.489
a really good answer. I was going to say something
00:56:10.489 --> 00:56:13.570
really boring, like it would be great if we could
00:56:13.570 --> 00:56:17.809
just use GitHub to update our entries. But what
00:56:17.809 --> 00:56:21.309
Johanna said is actually really a lot more important.
00:56:22.349 --> 00:56:25.690
this is a lot of work and usually when we are
00:56:25.690 --> 00:56:28.429
maintainer, we're doing the work by ourselves.
00:56:29.929 --> 00:56:34.170
But having specific groups or having more volunteers
00:56:34.170 --> 00:56:38.849
who are happy to even maybe also just work with
00:56:38.849 --> 00:56:42.929
the main maintainer in testing out the tool from
00:56:42.929 --> 00:56:47.230
their perspective and having maybe a meeting
00:56:47.230 --> 00:56:50.000
or a brainstorm or whatever in terms of What
00:56:50.000 --> 00:56:53.800
did you use it for? At the moment, for example,
00:56:53.880 --> 00:56:57.860
I'm writing a guide for Epiphy, which has a huge
00:56:57.860 --> 00:57:01.360
wealth of possibilities, and I cannot write about
00:57:01.360 --> 00:57:05.539
all of them. And it would be really great to
00:57:05.539 --> 00:57:08.639
have more people come on board and help us. And
00:57:08.639 --> 00:57:11.440
they don't need to take responsibility for their
00:57:11.440 --> 00:57:17.000
own tool right away. You don't need to be an
00:57:17.000 --> 00:57:19.969
expert. Um, if you're interested in tools, if
00:57:19.969 --> 00:57:22.650
you're interested in making tools available to
00:57:22.650 --> 00:57:26.070
researchers and the public, um, you're very welcome
00:57:26.070 --> 00:57:29.670
to come, come join us, uh, whatever your, um,
00:57:29.730 --> 00:57:32.920
whatever your level of expertise is. As long
00:57:32.920 --> 00:57:35.039
as you're kind of willing to put in some, some
00:57:35.039 --> 00:57:38.539
work, of course, but yeah, it would be really
00:57:38.539 --> 00:57:40.099
great. Thank you. I know that was a really good
00:57:40.099 --> 00:57:46.150
point. I agree with both Johanna and Martin,
00:57:46.530 --> 00:57:48.889
basically investing in the long -term development
00:57:48.889 --> 00:57:52.269
of the toolkit and bringing more people because
00:57:52.269 --> 00:57:56.710
this is an important project and it should evolve
00:57:56.710 --> 00:58:01.489
keeping in mind the need of researchers in today's
00:58:01.489 --> 00:58:06.889
time. Very good answers. I would just have been
00:58:06.889 --> 00:58:09.829
like, yeah, I want my category to be the biggest,
00:58:10.090 --> 00:58:12.750
I want big titles. But you all went diplomatic
00:58:12.750 --> 00:58:16.429
and I liked it. Thank you so, so much for coming
00:58:16.429 --> 00:58:18.869
in to talk today and thank you so much for your
00:58:18.869 --> 00:58:21.989
work on the toolkit as well. It is amazing resource,
00:58:22.409 --> 00:58:25.809
loved by many, used by many, and hopefully used
00:58:25.809 --> 00:58:28.150
by the people that have attended today as well
00:58:28.150 --> 00:58:30.670
and listening to this. Please do go check it
00:58:30.670 --> 00:58:34.460
out. The link to the toolkit is in the chat and
00:58:34.460 --> 00:58:36.179
it will also be in the description as podcast
00:58:36.179 --> 00:58:39.199
and we will be back in two weeks time with another
00:58:39.199 --> 00:58:43.320
stage talk but for now thank you to all of our
00:58:43.320 --> 00:58:46.739
speakers today and we will see you soon thank
00:58:46.739 --> 00:58:49.619
you all thank you for listening to the stage
00:58:49.619 --> 00:58:52.860
talk if you'd like to catch a stage talk live
00:58:52.860 --> 00:58:55.920
where you can ask the guest questions join the
00:58:55.920 --> 00:59:00.860
bellingcat discord server by visiting https://discord.gg/bellingcat
00:59:03.219 --> 00:59:07.199
The music you've heard is titled Dawn by Newer
00:59:07.199 --> 00:59:09.699
Self and is courtesy of Artlist.