The following are the outputs of the captioning taken during an IGF intervention. Although it is largely accurate, in some cases it may be incomplete or inaccurate due to inaudible passages or transcription errors. It is posted as an aid, but should not be treated as an authoritative record.
***
>> Good morning, gootdz afternoon and good evening to all
those joining us online. This work is shop is called Decolonise
Digital Rights For a Globally Inclusive Future. Before we begin, I
would like to encourage on-site and remote participants the scan QR
code on the screen here. And you know, the link is being published
on the Zoom right now. To express your expectation for the
session. And as a reminder I would also like to request that all
speakers and audience who we ask questions during the question and
answer round to please speak clearly and at a very reasonable pace.
I would like to also request that everyone participate and
maintain a respectful and inclusive environment in the room or in
the chat. For those who wish to request questions during the
question and answer sessions please raise your hand and once I call
upon you if on-site please take the microphone the left or the
right side, clearly state your name, the country come from and then
you can go ahead and ask the question. Additionally please make
sure that all mics are muted. And all the devices -- other audio
devices are also muted just to avoid disruptions. If you have any
questions or comment, or would like the moderator to read your
questions online put them in the Zoom chat and please start and end
your session with a question mark to indicate whether it's a
question or a comment. Thank you. We may now begin our session.
So thank you for joining the session with us online Orion site. It
is going to be a very popular concession that is going to delve
into decolonization of the Internet. I'm Man Hei Connie Siu, we
have Neli who is going to be the moderator and we have Keolo, going
to be the photographer for the session. Sod today we're going to
have the truth where the Internet is a space wherever one is not
always equal. Very far from being a level playing field. Instead
it reflects and perpetuates historical bias and power imbalances,
traditional marginalized groups create the -- of technology.
Results to digital colonialism and the dominance of privileged
groups in shaping technology design often leads the production of
off balance power dynamics. This has far reaching consequences as
digital content and platforms produced in the Global North continue
to be consumed by the Global South. Content from non-English
content regardless of the magnitude of hate or harm.
The unequal response to these strategies however further
highlights the disparate while features such as safety check and on
one click option to over pictures with the flag of country on
undergoing distress was -- after attacks in France, bombings in
Lebanon, failed to trigger a similar response. Where platforms
have also introduced fact checking measures for major elections in
the west, misinformation and disinformation bears some platforms
continue to plague the Global South. However, underrepresentation
of authors of color onion line knowledge platforms payments a stark
picture of the inequalities that possessed. Even voice assistance
designed to assist and interact with users have been found to
reinforce gender biases, normalize sexual harassment and perp pet
2008 behavior patterns on women and girls, this limits and puts
them in the forefront of biases. Hate speech marginalized
communities continue to reach online creating an unsafe environment
for those net Global South and those for the marginalized
communities. Uses in the Global South have the right to feel safe
and feel the same as users in the Global North. In this workshop
we are going to delve into the concept of decolonialization in
relation to Internet, and rights and freedoms onlines. Our
panelists will be joining us two on site and two panisms as well.
They will unpack the evidence that exists of gender stereotypes,
linguistic bias and racial injustice that are coded into
technology. They will shed light on how apps are often built based
on creators' opinions of the average user should or should not
defer. Further more they will also offer recommendations of how
online knowledge can be decentralized and ideological influences
can be dealing with the digital arena. They will pose practices
that can help decolonize the Internet and transform it into a truly
global space. Throughout the sessions we will address three policy
questions. One is that what are the colonial manifestations of
technology such as language, gender, media, and artificial
intelligence. And you know, that are on the Internet. Two is how
do we address these legacies that shape the Internet, and have
become the ongoing colonialism and determines its future. How does
the decolonializing Internet look like? What roles that is -- what
role do different stakeholders play in the process of
decolonializing the Internet, technology and digital arena as a
whole, how can we include marginalized communities in these
discussions? We hope by attending participants can get an
understanding in the context of decolonialization in relation to
the Internet and learn to recognize the ways in which bias is built
in technology and understand it is not fruly mutual as we think it
is.
They will also discover the algorithms mainly opinions written
into code, drawing data from actors, beliefs and systems that
perpetuates stereotypes and historical prejudice. To this session
we hope to aim -- hope and aim to have a conversation how really to
ensure that we decolonialize technology in digital space and pave
the way for a more inclusive future. Invite you to engage, send
questions and comments online and I don't have line. I will begin
by introducing the speakers. We have joining us, Jonas Valente, a
post doctorate researcher at Oxford Internet institute from the
University of Oxford and he's responsible for the co-lead of the
cloud network project, inside the Farrah foundation. He was
awarded a doctorate in sociology from the University of Brazil ya
in 2019. He is also a program -- we have -- then we have shall
Illini, who is a program director at lead on. She provides sup
fort to fact-checkers, newsroom and involved in addressing this
information. She's also the co-founder of Harav, India's only
independent digital news network. And one of the next speakers
Ananya who's here with us in person. Youth council over the recent
years, Ananya has been active in the global digital development
forum and has also been a next generation icon ambassador -- ICAN
over the 64, and ICANN 76 and holds a master's degree in
development and labor studies from Jawahala University New Delhi.
This is Ananya.
We have Pedro, joining us online as well. Innovation lawyer
at system industry, a Ph.D. student at UFBR, with an LMuscular
Dystrophy, from the University of Qiv uimba, a board member of IUD
A, as of bring sil and an organizer of the youth LAIF. We have man
dah, who's here with us in person. And he's from GIZ and a tech
advisory lawyer from Kenya. He heads the data governance division
at the DTC, Kenya and previously he worked as a data protection
advisor at GIZ. Also serves as a secretary of the Kenya privacy
professional association.
And with that we begin the session today. We will start with
Jonas joining us online to have a brief presentation. Yes. Have
you with us?
>> JONAS VALENTE: Yes. Can we get -- can I get the
possibility of sharing the screen?
>> MAN HEI CONNIE SIU: Yes. Please. Let me see that.
We can see your screen now.
>> JONAS VALENTE: Good afternoon. Good morning for me in
London. And even more for Pedro in Brazil. An honor for us from
the project to join this panel. I'm going to talk about the labor
conditions in AAI global prediction networks. This is super
important because normally we look in digital rights community to
the effects of technologies like AI, but we need to look also to
the workers who who are producing that. So first assumption is
that AI development and deployment is super dependent on human
labor. Unfortunately this human labor is characterized by a set of
features that make it very precarious and with very let's say
insufficient arrangements regarding a set of conditions like pay,
management, and collectivization. When we talk about data work, we
talk about activities like collection, curation, and notation,
validation. And throughout all this chain you have human labor.
So when we talk about this intelligence, it's important to know
that it's not so artificial. We need like thousands of work r
workers, and those thousands of workers are distributed all around
the world. But this distribution is not random or mutual. This
distribution express the legacies of colonialism when we have big
companies in the global market who are hiring and developing this
technologies and a workforce mainly in the Global South. We can
see here how the main countries are India, Bangladesh, Pakistan.
We also have a workforce in the United States, or the United
Kingdom. But mainly Global South countries are taking part in this
through business process, outsourcing, or digital labor platforms.
The fair work project assesses the labor platforms against the
set of principles. And we try to address the risks of platform
work and the platform economy. Which risks are those? Low pay.
How micro workers earned around $2 an hour and other reports and
studies show the same. So of course when we are talking about some
countries considering the currency, this may be not so bad but what
the studies are showing is that those payments, structures and
payment amounts, they are super insufficient to ensure like
adequate and meaningful livelihoods.
Another problem is the excessive overwork and job strain.
This leads to health issues. We have workers working 15, 16 hours.
Normally workers need to stwich day by night but thes need to be
awake during the Global North time instead of being awake in their
own country time. And this leads to exhaustion, leads to problems
related to sleep and very many other mental health questions we
have been finding in our studies.
Also workers suffer with short term contracts and precarity.
So normally if you have a business process outsourcing, you have a
one month or two month contract. When we mention platforms, you
don't have a contract in a traditional sense and these workers need
to search for tests all the time. Our 2022 report showed those
workers worked eight hours on unpaid tasks. And once again this is
a legacy that we see of colonial and capitalist regimes and work
arguments. Those workers suffer from unfair management and
discrimination. You can see this discrimination based on gender,
based on race, and ethnicity, and based on geography. We can see
the legacies of colonialism. Also data workers, they face
depersonalized ability and thur subject to extreme surveillance.
And finally, another risk of the lack of collective power.
And of course that this turns into more asymmetries between workers
and platforms. The fair work project is working across all over
the world, almost 40 countries. It's coordinated by the Oxford
Internet institute and vis-a-vis institute in Berlin and connected
to the German government. We are assessing location based
platforms, platforms in AI, and we have this five principles,
pay -- sorry about that -- we have these $5 principles, pay
conditions, contract managements and representations. We collect
data from different sources. And we run platforms to finish our AI
project --
>> MAN HEI CONNIE SIU: Please round up.
>> JONAS VALENTE: We are assessing specific AI companies and
we try to show the platform economy can be different and to be
different is part of the decolonizing process of AEI technologies.
>> MAN HEI CONNIE SIU: Thank you that was insightful with
data to back it up we could look at the fact that these are
concerning issues when it comes to the decolonialization of the
Internet. Another five minute presentation from another of our
online speakers, Shalini. Please go ahead and share your
presentation.
>> SHALINI JOSHI: Thank you. I don't have a presentation but
I made some points for discussion today.
Thank you very much to IGF. Thank you to the organizers of
this workshop. It's a real honor to be here.
I'm going to talk about the problems with AI in terms of
gender, in terms of language. And I'm also going to talk about the
work that we've done, the organization that I work with, has been
doing in order to address some of these issues.
So as we all know that there have been experiments that have
been carried out with generative AI, on how different image
generators visualize different people from different countries and
cultures. And when we look at these images, they almost always
promote biases and stereotypes related to those countries and
cultures.
When text to image models were prompted to create
representations of workers, for high paying jobs and low paying
jobs, high paying jobs were dominated by subjects with lighter skin
tones and were mostly male dominated.
Images that we see don't represent the complexity and the
hetero generality and diversity of many cultures. And people.
We also know that AI models have inherent biases that are
representative of the data sets that they are trained on. Image
generators are being used for several applications. And many
industries. And even in tools that have been designed to make
forensic sketches of crime suspects. And this can cause real harm.
A lot of the models that are used tend to assume a western context.
And the AI systems look for patterns in data on which they are
trained. Often looking at trends that are more dominant.
And they are also designed to mimic what has come before, not
create diversity.
So we're talking about inclusivity in technology. How do we
ensure that AI technology is fair and representative? Especially
as more and more of us start using AI, for the work that we are
doing.
Any technical solutions to solve for such bias would likely
have to start with the training data that is being used.
And to seek transparency from AI systems and from the
companies that are involved is also really important. Because very
often these companies are very secretive about the data that they
use to train their systems.
There's also the issue of language. Often AI models are
trained with data that uses mainstream languages. Often these are
languages of the colonizers, many AI based translation services use
only major languages. Overlooking hundreds of lesser known
languages. And some of these are not even lesser known languages.
So languages such as Hindi and Bengali and swau hee li which are
spoken a lot by people and by many people.
They also need more resources to develop AI solutions. And
from a social cultural standpoint, preserving these languages is
vital. Since they hold unique traditions, knowledge, and an entire
culture's identity while protecting their richness and language
diversity.
So in this context what is it that we are doing at MIDRAN, the
organization that I work with? We are a technology nonprofit over
the last ten years, as the Internet has evolved and changed chk
Idan has maintained a unique position as a trusted partner and
collaborator working both with Civil Society organizations, and
with technology companies. That harn nest the affordances of
digital technology to communicate. Our approach has been
consistent, we build collaborations, we build networks and we build
digital tools that make it easier for hyper local community
perspectives to be integrated into how global information
challenges are met.
We understand that our ability to work across community
technology, and policy stakeholders is a privilege and this is our
unique contribution. We see ourselves as facilitators and enable
ls of change. And we do this by developing open source software
that incorporates the state of the art ML, and AI technologies, by
building coalitions, a lot of these coalitions are built around
large events such as elections.
That enables skills sharing and capacity building. And this
multi pronged approach helps collaboration and hyper kablt
perspectives --
>> MAN HEI CONNIE SIU: Thank you so much for that, Shalini.
That was quite insightful to learn about the work that you do and
how the methods and codes and technology and that Internet that has
existed for as long as we have been raised on the Internet. If we
don't tackle them, talk about them, if we don't even realize these
stereotypes, gender biases are coded into our Internet and the way
we use dej tal technologies, we have a long way to go before
decriminalizing the ipt net. We're going to take another five
minute presentation or speech from another of our speakers, this
one is on-site. Before we do that I would like to share some of
the comments that we made about the expectations of the session.
We see that people expected reflections, candid direction,
articulation, radical, honest manifestations. Of course the link
is still on the Zoom chat. If you would like to include your
expectations you may still go ahead and make the comment.
And Ananya, you may go ahead please.
>> ANANYA SINGH: Thank you so much. Let me begin by saying
that I'm very happy to be here in gentleman is pan-. And no, its a
he not just because Japan is such a beautiful country and the
people here are so nice, I mean well -- of course they are. But
also because I can finally live a day where I do not get spam bid
calls from a range of companies trying to sell me their products, a
bunch of coaching centers trying to send me to the engineering
institutions with the aid of their tutors. By the way I have a
master's degree in development studies. So engineering was clearly
never my choice. Random calls into agents forcing my to invest in
certain deals or just another customer is you support automated
call trying to vie for my attention from work. One question that
always comes to my mind when my phone rings and the true caller app
detects it as a spam call is how did they get my number? Who gave
them my number? And why did they give it to them?
Why was I not asked? Given that it is my number, and my
number is connected to very obviously a ton of different data
related to me and since I own both the number and any data related
to that number, I should have been asked. But I wasn't. And I'm
sure we are all very familiar with those lottery means, we have a
dedicated spam folder where all those gait deals an and gone in a
day bumper offers and their lines of ads even keep lurking. How
did they choose you or me? I mean I have never been that lucky in
my entire life by the way.
So who give them out our email addresses and if they found
them are they going to be very far from our residential addresses
or bank account numbers? So the way we live our lives has become
excessively dependent on virtual and online activities and even
more so after the pandemic. For instance social media, GPS, health
labs, taxi apps, Google searches, all of them require and p access
to our personal data, details set to public or private are
available for usage by online companies. The principle actors here
capture our everyday social acts, translate them into quantifiable
data, chk which is analyzed and used for 9 gin racial of profit.
In the book, because of connection, the author's Nick holdry
anulusese may dra emphasize that instead of natural resources and
labor what is now being appropriated is human life through its
conversion into data. Meaning our online identities have become a
commodity which can be exploited and used for capital gains,
controlling our time and usage and influencing important decisions
or processes in our lives. Hins the term, data colonialism.
But I know some people do contest the usage of the term data
colonialism because historically colonialism is unthinkable without
violence, take over of lands and population ls by physical force.
That's true. Let's take the exams of these requires of the demand
document. It was made to inform the natives of the colonialists
right to conquest. Con keeft doors demanded the new conditions in
Spanish. Which no local understood. Think of the terms of service
we sign up to every time we join a platform. Unclear, long, full
of jargons which rarely have the time to read so automatically like
a reflex we click on "I agree" do we really agree? Unknowingly we
are giving consent to being dragged online, being called at all
hours to be sold insurance policies for the children. By the way I
don't have is yes. And hence our ignorance, implied or uninformed
consent for these kinds of data collection provides a very valuable
yet free raw material, data. Once a single fisher from a famous
company stated that data is more like the sunlight than oil.
Meaning a resource that can be harvested sustainably for the
benefit of humanity. But this very idea makes my personal data a
nonexcludable natural resource available for public use. But does
it not contradict the very what personal in personal data? I'll
leave you with that. (Word).
>> MAN HEI CONNIE SIU: She's the only person who's been on
time since this session started. Thank you. Thank you very much
for that. Ananya. We're going to take a five minute as well
presentation from Pedro who's joining us online. Pedro are you I
don't know line?
>> PEDRO DE PERDIGÃO LANA: Can you hear me? That's great.
Everyone, I hope you are all well. And with the preholiday morning
here in Brazil.
But to get to the presentation what I want to comment on with
you today during the session, is just -- let me pull the time here.
There we go.
Is the results of a research project by the focussed on youth
named glittry 2.0. Ifs an amazing problem with many interesting
and diverse spaces and I recommend you all seek my permission about
it as a way to repeat in the regions. And for the sake of time,
back to the real content of the presentation.
It is simple, you are -- per mutation, revolution as a
reaction and the team I tried to interpret our research, intellect
tulg property. So governmental regulation is probably one of the
most important threats we have to the Internet when we are talking
specifically about the dangers of presentation. But it's important
to see are what is behind this regulatory propose VIII. To be more
precise, what serves us justification for these movements, the
agomented that I will try to put forward here is that even when
this is not the real reason that motivates political authorities,
especially authoritarian ones -- triktly related by distinctions
originated in digital colonialism. Be it from nation or tech
companies or countries who have much more steering power modeling
the Internet than others. Even if that's not implemented in such a
direct and explicit manner. We can see those in these larger
companieses, and expanding the legal systems of their home
countries to every corner of the globe imposing digitalization 6789
even though it doesn't follow the standards of the nation laws that
actually apply.
This is where intellectual property comes in. The digital
millennium act or DMC, the copyright reform for digital society in
USA. Establishes systems of -- and codification. That are similar
to the rights offered, the copyrights order and the largest content
based platform seems to have repeated those systems all over the
planet. Sometimes of course with great support from the
international lobby of the American entertainment industry. And
similarly, when I go to Brazil pay for example that responds to
allegations of copyright infringement on this content based
platforms, I almost always see explanations on how pay use works.
An institute that simply doesn't exist in the Brazilian legal
system, this is an exceptions for permitted users of copyright
boards. Of course this example may seem strange to some. So how
many people actually care about intellectual property when on the
search, information or freedom of expression? But apart from the
fact that this -- all these areas are linked, and Brazil for
example even have a -- something like a bill that intended to
create obstacle for -- through copyright mechanisms.
The most important point here is just to exemplify a broader
behavior that affects a lot of -- and maybe instrumentalized by
unintended actors. If a platform doesn't care about conveying an
image that will follow as simple today as sovereignty as national
legislation, you can only imagine what a place exists for movements
that want to showcase the national interactions made possible
through the Internet as something dangerous or something that needs
to be controlled.
Internet content and platform diversification, talking about
user experience, language accessibility et cetera. Is not the same
as augmentation. Not only that, it's not just only the same, but
this diversification, to actually set in context mayly be an
important -- to revelation that may result in fragmentation.
So back to you.
>> MAN HEI CONNIE SIU: Thank you. Thank you so much for
that, Pedro. That was quite insightful. We' take our last opening
remark from Tevin, frl GIZ Kenya.
Yes, it's working now. Go ahead.
>> TEVIN GITONGO: Good afternoon everyone. My name is Tevin
Gitongo. I think we've had quite a number of presentations. And
mine is going to take a different tangent. Mine is going to show
you how we are trying to decolonize the digital future. You've
heard all the things that are happening and some sound scary. Ours
is more let's try and actually solve it. Let's put our money where
our mouth is. I'm going to make a short presentation of the
project we're working on as Jzit, Kenya, an information center,
support the by the German government working together with Kenyan
government. And in our own little way we aren't perfect but we are
trying to see how we can do this with different aspects. We
recognize developed presentations on AI, but when the digital
rights future it's not only AI only, it has to be every other facet
as well at that builds up to the AI. That's what we are trying to
do in our own small way.
So the project as you can see the objective to support digital
transition towards a sustainable and human-centered digital
economy.
And I'm going to look at two -- three visions and missions.
But I'm going to look at two major ones that affect this panel.
The first one is recognize that we must make technology work for
people and throughout the presentations you've had, that's
something going wrong particularly in developing countries. The
technology being made at some point is not working as ideally as it
was intended. Rights based on democratic digital society. We have
to be aware of that.
And so what approach did we decide to take with this? I can
say interesting experiment. It is on one hand to leap frog in
digital economy. We decided the first thing we are going to do --
working together, to support the local judicial innovation
ecosystem. To build capacity on digital security. Forced a driven
economy and work towards a job economy and digital economy. All
this actually build up together. To enable that. Other thing that
we've done is to build Kenya's digital society. This is exploring
imagine like AI, doing a lot of work there. And it can be
socioeconomic benefits. Let me show you an example. Assisting pun
services but user centric way so we don't use anyone behind and
building capacities on data protection. And also focus on bridging
digital defines, no one is left behind. The youth and persons with
disabilities.
So at what the approach we took is as you can see there, on
the side there all our stakeholders, literally the perfect example
of IGF, in practice and in working, in everyday work that we do.
At any one moment like in my work I deal with all those
stakeholders. Because we recognize that fact. One of the best
ways to actually achieve a future digital rights would be you leave
no one behind. We have governments, in our teams, private sector,
Civil Society. And academia.
So what has been the impact and achievement so far? I'll
highlight a few major ones. There a quite a number. Ones that
as -- to this.
So the first one was study on human centered Civil Society
approach in the Kenyan center. We are known as a fin tech
powerhouse in terms of the work that we do there. But out of that
we've also started creating some level of biases, some of the
things we've been doing is how do you leave no one behind in this
sector? The other thing that we've done is data protection and
privacy from a gender perspective. I think that's important
because we always forget that the most vulnerable groups,
particularly when comes to data protection, in most cases are
women. We decided to look at data protection and from that
perspective, how to enable participation online.
The next thing that we did was...I'm going to jump to our
other -- yes.
Strengthening -- sorry.
Strengthening gig workers rights. Every year we publish a
report woor we rank digital labor platforms and the principles and
how are they performing.
And the other one comes to AI and leaving no one behind. One
that I'm excited about is building local solutions. And other
things that we did for example, working with Kenyan actually --
Kenyan entrepreneurs and Kenyan coders we are creating chat bots.
The versions that you see of open AI. This one is locally created.
Able to speak English, swau hee li, a version of English and swa
heelly. In that way some of these products are created are kind of
geared towards the persons and they're able to help.
So let's just -- and also English and we developed the
first -- accessibility standards. Just some of the few ways that
we are trying to I can say decolonize digital rights and an
overview. Thank you very much.
>> MAN HEI CONNIE SIU: Thank you very much for that, Tevin.
I think our collective thoughts are always very well needed in
these kind of issues. Our panelists have shed light on the concept
of decolonialization in relation to Internet technology and human
rights and freedoms. It's time that we engage in discussions that
go goes deeper into these concepts and objectives how we can
harness these innovations and issues to responsibly create
something more sustainable and equitable for inclusive -- global
inclusive digital future. I would now like to ask -- we will start
with Jonas, who is online. Jonas, what are some of the ways in
which cheap label from the Global South powers contemporary digital
products and services?
>> JONAS VALENTE: Labor is key for all AIdevelopment. And
this is why lots of companies are using digital labor platforms,
because digital labor platforms circumvent the social protections
and digital labor rights. Basically digital labor rights.
Sometimes talking about the 19th century rights like minimum wage
or freedom of association.
And using that, those companies can benefit from this cheap
labor. And those workers unfortunately are not being compensated.
Do not have health and safety protection measures and don't have
the rights that we talk about as once again, from the 19th century
to the 20th century. Unfortunately this is becoming a role in the
data services global value chains, including AI. And that's why we
need to address this issue and talk about how to ensure those labor
rights to workers. All around the world. But focussing
specifically on what's happening in the global south.
>> MAN HEI CONNIE SIU: Thank you Jonas. I have a question
for Sally. Why are these conditions so bad and how is the project
working to improve them? Jonas you have the floor. It's to you.
How is the project that you're working on working to improve them?
>> JONAS VALENTE: They are only addressing on location
platforms.
>> MAN HEI CONNIE SIU: Okay, we're talking about the Internet
Governance, we have ipt net issues online. We're going to go ahead
and move to Shalini since there's Internet blockage over there.
You mentioned some of the work you do at mean na, during your
opening remarks. What forms does online hate and falsehood take
while preding in the AC region?
>> SHALINI JOSHI: Thanks. I'm going to focus on the issue of
gender in the Asia-Pacific Group region and focus on South Asia.
So women, trans people, nonbinary people in South Asia are
regularly tear getted with online disinformation. And this
disinformation is propagated in an attempt to silence already
marginalized individuals and make it difficult for them to safely
participate in public discourse.
Much of the work on gender disinformation covers women in
politics. And those in the public domain.
Research also shows the narrow definitions of gender
disinformation and the carbon -- have been focussed on women public
figures are sometimes side lining affected girls and women and
gender minorities who do not have a public presence.
And jendzer disinformation as we know can take many forms that
includes hate speech, intentionally misleading information, rumors,
attacks on the character and affiliations of people and attacks on
private and public lives of people.
Which impacts people in a way that they are either self
censoring or removing their social media contents or living in
hiding. There are direct and indirect threats to their lives. And
also generally enforcing stereotypes of vulnerability.
So what we're trying to do at Miran, is we are developing a
data set on instances of gender disinformation to build more
evidence for supporting research and policy action. And we have
brought together a diverse set of stakeholder groups in south Ashe
to work collaboratively to define gender disinformation from a
value ofation perspective. To identify, document and annotate a
high quality data set of gender disinformation and hate in online
spaces. For better understanding and countering the issue.
We're going to use machine learning techniques in the process.
And as we document more instances of gender disinformation online,
we feel that the technology that we use will become better at
locating additional content and thereby creating a working
psyching.
>> MAN HEI CONNIE SIU: Thank you Shalini. Thank you for
that. When you started answering the question I was going to make
a follow-up question about some of the best practices and measures
that you guys have taken in place, put in place rather to counter
online hate that targets marginalized communities and with regards
to your context, you're talking about women. But you answered that
when you're talking about the data set that you guys are
developing. So thank you for that.
And Ananya, you talked about when you were making opening
remarks, you talked about data, a lot about data and how it's
really affected. It's the key. So what are some of the
implications of data colonialism and surveillance and on digital
rights uflt. How can individuals and communities really reclaim
control over their personal data that they time sometimes aren't
even aware this they've given out? And how do they protect their
privacy in the digital realm?
>> ANANYA SINGH: Yes, apparently it's no longer oil but it's
sunlight.
Well, historically the era of colonialism ushered in by folks
that came to that new world to expand empires through
infrastructure building and precious metals extraction. Now like
every other thing colonialism is also going digital. Establishes
extensive communication networks like social media and harvests the
data of millions to influence things as simple as advertising and
as critical as elections. Data colonialism justifies what it does
as an advance in scientific knowledge, personalized mafshging, or
rational management, just as historical colonialism claimed itself
to be a civilizing mission. But yes, some things have changed. If
historical colonialism and -- there are resources and the bodies
that worked on them. Data colonialism's power grab clusters around
the capture and control of human life itself through appropriating
the data that can be extracted from it for profit. Data
colonialism is global. Playing out in both the Global North and
the Global South dominated by powerful sources in both the cache
east and the West. Unfortunately regardless of who directs these
practices or where these practices take place they often lead to
the erosion of privacy rights as such individual's personal data
collected and analyzed and used without their knowledge or explicit
or informed consent. Chk. Like you saw in the example that I gave
you about the spam calls I get, there is little to no redress or
mechanism. Yes I can block and report. But can I happily live
ever after? No, because I will get another company which is
actually, you know, employ another spammer waiting to call me again
to sell their policies.
My data, your data, and now in the hands of so many people
that it is going to be extremely difficult for us to individually
trace and then erase our data. Hence this will ultimately result
in a loss of autonomy and control over our own personal
information. While our data may be widely dispersed, the power to
capture and control our data continues to remain concentrated in
the hands of a few. This can lead to a lack of transparency,
accountability and democratic controls over data practices,
potentially undermining individuals' rights and freedoms: The
collection and analysis of personal data it perpetuate existing
inequalities like some of my able panelists have mentioned.
Training technology on biased data can lead to biases in
algorithms, unfair targeting, discrimination and the list goes on.
These practices can be used to manipulate and influence
individuals' behaviors, opinions, threatening individuals and dmok
sis. We have seen that happening already. Undeniably ideologies
such as connection, building communes, personalize ation will keep
incentivizing corporations to collect more of our personal data.
Hence the only way to prevent data colonialism from further
expanding is to question these very ideologies. Individually we
must prioritize data minimization like be mindful of the
information we share online or limit the amount of personal data we
share with technology platforms. I personally do this by limiting
my social media presence which is very good for your mental health
as well. I like to call this digital minimalism. Further, think
twice before you agree to their terms and conditions. While it is
easy to be fatigued by the almost inxree henceably long document
written in complicated language, take time to think before giving
into impulse of clicking on "I agree" so I stop with that because I
don't want to take more time than I have been allocated. Thank
you.
>> MAN HEI CONNIE SIU: Thank you for that Ananya. That's
insightful. I do have comment to make honestly. We want people to
be able to be at ease, comfortable, be safe on the Internet, not
have to restrict themselves from using the Internet or social
media. This is something that we have to talk about again in
another session or towards the end of this session about how we
also have to talk about data, making sure that data is utilized
appropriately -- with purpose not just for spam calls like you
experience. I will move to Pedro who ace online. My question to
you is do multinational platforms care about the legal and cultural
particularities of the countries in which they operate?
>> PEDRO DE PERDIGÃO LANA: I will try to shorten up my
presentation so we can give the floor back to Jonas at the end of
the section. I don't think they care, if you -- in particular the
market in which we are trying to sell your services. Usually less
profits or at least more cost. But this concern just go as far as
they make the costs and adaptation can be considered not too high.
This is a problem when they consider the difficulty of measuring
the indirect and internal costs that platform will serve in the
fragmentation scenario. For example platform is investigated in
the research project translated the pages about intellectual policy
properties, but whether you browse for more details, no, sir
something as simple as the translation of some pains were normally
done or the hyper lippings, letters to English versions. One of
them which was not content based had only the most basic page
translated.
>> MAN HEI CONNIE SIU: And how is this reflected in the
global hup rights system as a rule it still has to serve off
national legal system that determine the factors of jurisdictional
conflicts, rather?
>> PEDRO DE PERDIGÃO LANA: I think this reflects directly on
human rights. Intellectual property is globally considered human
rights. But here is that we had some international frameworks,
human rights are not interpreted the same worldwide. So freedom of
expression is a good example. Some countries see it as a much
broader right than others. Copyright may be stronger or weaker
when presented with rights such as education or access to future.
If that terms the policy z around such concepts they should at
least do it in a way that is not so clearly unbalanced to the civil
perspective. Especially saying they should -- quite frankly a bit
offensive since it really wouldn't cost that much to get someone to
do a quick review on the legal policies, deliver some adaptation,
even if fishl. The problem here is this image that those platforms
simply do not carry some basic elements of societies that they have
as markets for their services and products. Especially when
they -- we see that they can evidently -- they can adapt. So as
one can observe, with changes made because of the general
legislation called nest DG, especially on social media.
>> MAN HEI CONNIE SIU: Thank you. I'll move on to Tevin
here. What jazzy is doing with the gods to welcome the community,
especially the marginalized communities. How can digital literacy
and digital skills training be reimagined in a way that is such
that it empowers the marginal I-autoed communities and bridges the
digital divide? And you know, in such a way that is ensures that
everyone has the necessary tools to fully participate in the
digital realm?
>> Thank you for the question. I think I'll pick up from the
question you asked earlier on do entities care about the legal and
cultural considerations? (Tevin Gitongo) the less shun is you have
to care about the cultural considerations to have impactful
trainings or digital skills. A case of you have to bring yourself
to the level and be there with the partner that you want to achieve
the training to. And maybe thinking of it practically, how do you
do that? How do you actually demonstrate that you are aware of the
person's context and how you can help them to bring them up to
where you want them to be in terms of listening the digital divide?
I think how I look at it -- I know I look at it as a four-step.
And the first one is the stakeholders that you work with. Because
more or less are not always guilty of working with stakeholders who
have no clue what's happening on the ground. You tell them you're
going to do this and they tell you yes you have a he done a
training and you realize this was the wrong stakeholder, they
clearly did not understand what was happening in this context.
That way you're really -- your training doesn't interest impact.
The next thing I look to is accessibility. And how I look at that
is in the relation to democratizing the knowledge. And by this I
say when you do a training, it should be one that you're actually
transferring knowledge, not just ticking a box. There's a huge
difference there. Because most cases we are ticking boxes. But
not actually transferring knowledge and knowledge that actually
helps them grow.
And one of the things we've done with that I'll give an
example of -- I see my colleague is also here. When you're
developing the AI chat box. Because it is a skill we're trying to
transfer. We brought Kenyan developers in the room. We brought
other developers, I think it was from Europe who have expertise in
developing such models, and you are like we want you guys to teach
them to teach each other. Not just to teach them. Teach each
other how to develop this. They're coming with Indigenous
knowledge of how to in swa hee li, and English. Maybe they come
with the knowledge of how to develop the systems. And what
happened is after developed the system, the next system that we're
developing, another one for the Kenyan's data protection
commissioner, the Kenyans running the show now. They're now
developing everything.
So you start seeing just slowly reducing that cup. The.
.
The next thing is affordability of course. If you really want
to create impact, you have to create training that people can --
goes back to accessibility and lastly inclusion of everyone. Than
can also be done practically and one of the things I think I
mentioned assisted developing is ICT standards for persons with
disability for Kenyan. When you're designing a system, how to
design it for persons with disabilities and you don't leave them.
Kenyans are advertising a lot but we are forgetting -- but I'll
leave there.
>> MAN HEI CONNIE SIU: Thank you so much for that Tevin. I
think that you know, with everything that all the panelists have
said, it always goes back to bridging the digital divide. Digital
skills. Making sure that people are aware of these things and they
know how to protect themselves, how to use it and they know what
the issues are and how to tackle them. When it comes to any matter
of internal governance, if you ask mere we shall' not going to go
any farther if we don't tackle the digital divide. Chk we're going
to go back to Jonas who had issues online. I think we have some
time we can spare. He's back now and he was going to tell us about
the ways in which cheap labor from the Global South and powers
digital products and services. Jonas can you tell us about why
these conditions are so bad and how is the fair work project
working on improving them?
>> JONAS VALENTE: Thank you so much. I'm sorry about my
connection problems. These conditions are bad because platforms
find -- found a way that my connection will -- I think I will
freeze again. I hope I don't.
Because platforms found a way of circumventing digital labor
and social protections. And by doing that, companies can hire
cheap labor and that's why we're seeing low pay, health and safety
issues. And management problems all around the world. Study has
estimated 163 million online workers, this is a representative
number of people. The fair work process assess that plan.
Platforms all around the world in those 38 countries. We analyze
and scored those platforms according to five principles, pay
conditions, contract management, and representation. In a scale
from 0 to 10. And we launch rankings. I invite all of you to
visit our website fair.work. You can see maybe platforms from your
country and check what they are doing or what they are not doing to
meet basic standards of fair work.
>> MAN HEI CONNIE SIU: Thank you for that. I would like to
thank all our speakers on-site and online for sharing their
insight, sharing their experiences and the efforts that they're
working on. And I would open the floor to four questions, both
on-site and online. I don't know if -- online do we have any
questions online? If you're on-site and you have a question you
may go to one of the standard mics. State your name, the country
you're from and go ahead with your question.
We have one question on-site.
>> AUDIENCE MEMBER: My name is Niturla, one of the youth
ambassadors from Internet society, from Italy. As the panelists
anticipated I am understanding there are a lot of stereotypes such
as specific legal diversities that are not always respected. Also
lack of accessibility. Also the need to respect privacy. And all
these different problems and needs are not always really respected.
And all of that is because of economic patterns and interests
worldwide. But some of them for example privacy, I would argue are
also global rights. We can discuss about also being then human
rights.
I would really be interested to see let's say a taxonomy of
specific local needs that are not respected by specific
technologies of the Global North. When it comes to culture,
history, or political characteristics.
So I would like also to understand which are shared also with
the Global North and which are not.
And with not, I mean not regarding people originally born in
the Global South that lately got to live in the Global North, but
specifically populations that plan to thrive in their own country
of origin.
So the idea is to understand which needs are local and which
are global. Thank you.
>> MAN HEI CONNIE SIU: Okay. Do you want Jonas to answer
that or it's open to any of the speakers? Sir? Okay. So Jonas,
do you want to take up this question?
>> JONAS VALENTE: I think that what I would like to say is
that when we talk about this national and cultural context, what
fair work project is bringing is that we have one very serious
problem that has been addressed here, by other speakers, that is
the biases, discrimination. That is faced by the users. But we
also need to consider what is behind the digital technology
prediction and that's why we highlight this discrimination. And
the consideration of the local context and for instance when Pedro
brings the discussion on national regulations we also need to
consider as well the national regulations about work and how those
national regulations, the national and local context, and the
different populations in the diversity of populations and cultural
expressions can be considered and -- in its own characteristics in
the Internet as a whole. But especially in digital labor platforms
and global platforms. And that's why I believe that this
discussion that Pedro brought and now that we have the conversation
needs to look to those diverse contexts and groups. And at the
same time think about how to incorporate them also not only in the
digital and data practices but in the regulatory efforts as well.
>> MAN HEI CONNIE SIU: Okay. I'm also going to invite Tevin
to address the question, very brief remark.
>> TEVIN GITONGO: So maybe you asked what are local and
international. International I will say privacy. Thank you.
Affects everyone. Doesn't matter whether it's global or not the
Global South. We see it in Kenya every day, we have the
commissioner's office and work with them. And the same issues that
are raised in European countries are in Kenya, and AI product, why
am I getting these marketing messages? How did you take my data?
Issues of consent. Where are you using it? It's been interesting,
where are you transferring the data to? They're asking questions
that you would find in -- this is not just a consider Kenyan, it's
rural Kenyan, you will go and talk to them and say okay I saw this
application, however they told me to do this. And I'm wondering
why they told me to do this. It's something that everyone is aware
of. In relation to local, I will say languages. Because when
you're developing example natural language processes, suddenly most
of them are geared towards Global North. However, English, the
pronunciation is very different. Language is being used. But we
start looking at local aspects of especially languages. Because
that's the only way you start brinling the digital divide, not
everyone can speak fluent English or fluent swau hee li. And you
need to --
>> MAN HEI CONNIE SIU: That is going to bring me to a
question for Pedro. Is the risk of regarding the search for
balance of hour r power relations between countries, is it a risk
and how does this affect the Internet as a global network?
>> PEDRO DE PERDIGÃO LANA: I would like to go over what was
just said by the previous speaker that I would use the same example
but invert it. I think language is an international issue.
Because even though we are adapt to each country, it's 9 same issue
that we have on the road. Privacy, you can interest interpretation
of privacy, what is not. And that's exactly what is especially
dangerous when you're talking about not diversifying what they are
doing and how they are -- do not do that in international -- they
prefer regions to others.
So in a period that international relations are becoming
increasingly tense and discourse against -- on the rise, it seems
very easy to expose the facts about how those relations work such
as talking about how these platforms may be expanding influence of
certain country or even acting directly on their behalf as we
learned with Snowden. But also easy to extrapolate the -- support.
International nature of the Internet as a problem in itself. So
doing those small things such as translating the content correctly,
international translation may be, exactly what we need to avoid
having an Internet, having the Internet as a -- severely affected
in an active way.
>> MAN HEI CONNIE SIU: Thank you. Thank you for that. We
have learned that we have some questions from the online
participants. And I would like to call on two, Neli, to request
the questions out loud for the audience. Neli, you may take the
floor, please. Are you with us, Neli?
Okay. It seems Neli is not with us. And...any other question
from the audience here, on-site participants?
Neli, we think you are muted, Neli.
Please unmute your mic and take the floor.
Technical, can you please give -- can you please help us give
the floor to Neli, please. Unmute her mic.
Okay. If there are no other questions and it seems we're
winding up this session today, I'm actually very thrilled again to
invite our speakers to share their invaluable recommendations to
the following.
What should -- (cross-talk).
>> MAN HEI CONNIE SIU: Was that Neli? Neli, is that you?
Okay. We're going to go ahead. If Neli happens to unmute her
mic we'll just take questions from her. But until then I'm going
to ask the panelists here who have shared their insights and
experiences for their recommendations regarding the following
questions.
What should decolonializing digital rights look like? Before
I give you the floor I would like to encourage the audience to
seize this opportunity to share your recommendations again by
scanning the QR code that's displayed, on the screen shortly. And
now I would like to welcome Ananya. Please go ahead. Tell us what
should decolonializing the Internet look like?
>> ANANYA SINGH: Well, let's say this. My blood group is B
positive. There you go. You have another one of my personal data
points. Anyway, being the positive person that I apparently am, I
believe that every cloud has a silver lining. So this cloud of
data colonialism presents an opportunity for us. An opportunity to
create ethical systems which run on the principles of A, ownership
by design where users are provided with clear and understandable
information about how their data will be collected, used and
processed. Shared, stored or erased. It involves obtaining
informed consent that is granular and specific, allowing
individuals to make inferred choices about will their data. B
minimization and ought MissAtion, only relevant envelopes is
collected and processed and wherever possible such date it is kept
anonymous and inkrichted. Reduces unauthorized access. C, there
should be an option to be forgotten or easily revoked consent when
desired. There are options to be forgotten but the option to
revoke consent has been a complicated process so far. D,
mechanisms for accountability and roo he dress in case of data
breaches or violations are hard to find. This involves providing
individuals with avenues to support their rights, seek any harms
and this should go beyond blocking and reporting accounts.
And E, I just want to finally take note of this. The whole
in#250i89ed attitude that makes data colonialism possible must be
done away with. Spelled simply, for example I was born with a
name, my name is a data point. Just because I provided my name to
my school on the day of enrollment does not automatically translate
into their unprecedented right over an unchecked use of my name for
the rest of their existence. Data used is not a right but a
permission slip. Data reuse is not an entitlement. But once
again, a permission slip. Thank you.
>> MAN HEI CONNIE SIU: Thank you. Thank you so much for that
Ananya. And I think we have access to Neli now. So we're going to
take a question from her online.
Neli, you may unmute your mic please and ask the question to
our panelists.
>> Neli. Many that I am you for letting me on an online mic.
Initially the question evolves according to your discussion is like
this. How can data -- kills training, be reimagined to empower
marginalize communities and bridge the digital divide and ensuring
that everyone has the necessary skills to fully participate in the
digital world?
Which of us --
>> MAN HEI CONNIE SIU: Can you please repeat the question
Neli?
>> NELI ODISHVILI: How can digital -- skills training the
reimagined to power marginalized communities and bridge the digital
divide, ensuring that whichever everyone has the necessary skills
to fully participate in the digital world?
>> MAN HEI CONNIE SIU: Okay. Tevin is going to take the
question.
>> ANANYA SINGH: Just to help Tevin answer the question, it
basically means how could we use or program structured digital
literacy programs which I assume will help people to better
navigate in a world in which is more decolonized. How could
digital -- the process of the Internet.
>> TEVIN GITONGO: I'm a proponent of -- and I think I'll
keep -- you have to bring yourself to the issues of the person.
I'm going to give a good example. We were having a discussion
recently. There is this in Kenya, we have this -- sell groceries.
They have these little shops where you go buy groceries. And we
were thinking how do you enable them for example use digital tools
to enable the sale of their products?
And the discussions we were having -- we used -- having so how
do you -- how do you go to them?
You know, because -- and not make them come to you. If that
makes sense. Because in most cases it's always ask, telling them
come. But now it's how do we go to them? How do you go to them at
their level and work with their skill? Because they already have a
lot of that skill. And just empower that. And I think that's what
the -- the challenge and discussion should be. So it's something
that we are also lining -- that example I just gave, it was just
something that we are trying as well. Like how do you go there,
how do you walk with them where they are?
And I can't say we have the complete answer to that. It's a
line in process. But I'm a big proponent of find people where they
are. Don't make them come to you. Because that's more pattern.
You look for them and walk with them from where they are. One of
the things that studied when we were talking to them was how much
knowledge they have. They do have a lot of that. One of them was
telling us you click on this and I don't know what I'm clicking on
but it doesn't make sense when I read it. And you just say in
terms and conditions, it's like 30 pages. You say aagree and move
on. But they are cognizant of the fact I'm giving away my data.
Perhaps it's coming to them and breaking it down to a point where
they also understand.
>> MAN HEI CONNIE SIU: I like that you mention that. In
digital development you have to design with the user. Because at
the end of the day if you're looking to benefit them and they need
to be actively involved in the process, you need to know what they
challenges are, what their perspectives are, what they think is
going to benefit them and include that in the process. And that's
very true.
I would like to continue the recommendations that we're
getting from our panelists. About what should decolonializing
digital rights look like. We will -- Ananya has given hers. We're
going to move on to Jonas online who is going to share his --
please note there is -- I'm sharing my screen and there's a QR code
where you're supposed to -- should be on the panel lock. On the
screen. You can scan it and I'll send the link in the chat for the
online participants as well to make their comments as well. So
Jonas, please share your recommendations on what decolonializing
digital rights should look like.
>> JONAS VALENTE: Thank you so much. I would say that
decolonialize digital technologies involve not only decolonialize
the use of digital technologies but also the prediction process.
That's why we need to incorporate the labor dimension to our
decolonialization agenda. And this means to ensure not only basic
standards of their work, it's why what we are assessing our
project. But a radically and structurally different work
arrangement where workers are not exploited, where we don't have
international, national, local and population groups, and
symmetries. And we have -- where workers are not exploited
anymore. So I believe that we need toin corporate this through our
agenda an and to quote a Latin American philosopher calleden Rika
Duso, it's also not about declon yalize. But to -- oppressed
people and create something radically new.
>> MAN HEI CONNIE SIU: Thank you for that Jonas. Shalini?
>> SHALINI JOSHI: I'm going to be very brief and say that in
order to decolonialize digital rights it's really important to look
at who is being included in the process of creating digital tools.
We have to involve at the local communities in involving -- in
creating data sets. Something that I talked about earlier as well.
We also have to make sure that there are people from marginalized
communities who are involved in analyzing the data, annotating the
data, in actually creating the technology because it's these people
who understand the context. The language, and the issues, much
more than, you know, technologists, and coders and developers
sitting somewhere else.
So involving the people in the creation of the technology,
making prols susses more inclusive. Ensuring that many, many
languages are being included. In the way that you know, we analyze
data. All of that is really important. Chk.
>> MAN HEI CONNIE SIU: Thank you for that. And Pedro? What
are your recommendations?
>> PEDRO DE PERDIGÃO LANA: Yeah. I already talked a bit on
my last comments, but just to be very brief, I think that should --
more to that local scenarios. And countries that are historically
more -- how the Internet is modelled, should actively try to share
these powers, these capacities. Just not about just decolonizing
the digital space but preserving the ipt net as we know it as a
global network and -- so that's it. Thanks for your attention.
>> MAN HEI CONNIE SIU: All right. Thank you. And we'll take
our last recommendation from our final pantist, Tevin.
>> TEVIN GITONGO: Praeps my recommendation would be to ask
ourselves four fundamental questions, who is developing the
systems. Second one is why are they doing it? And most cases it's
for economic gain. If you're being honest, the baseline of this
whole conversation is economic gain. And what they stand to
benefit. As Ananya said there's a new sun and the politics of
data. Everyone wants to be the ruler of data now. Second is where
are they being develop the? Where are they marginalized. Why it's
being developed. Someone sitting in sill Kahn valley. The last
person -- because of where they are. The last thing is what is it
for? By the end of the day. Yeah. Thank you very much.
>> MAN HEI CONNIE SIU: Yes, that's very true. Chk and I
think all our panelists have shared very thought provoking and
insightful experiences. And insightful expertise on this topic.
As we conclude this session today I'd also like to express my
gratitude towards online and on-site panelists for their expertise
and thought provoking contributions. You have been very
instrumental to deepening the understanding of complexities that
surround the decolonialization of data and technology. I'd like to
thank the audience, on-site and online for engagement and questions
and for being here today. Your participations have enriched our
discussions. In closing I would like us to remember that the
journey towards a decolonialized ipt net and digital landscape,
it's ongoing. It's not static. It's not something that's already
established. It's ongoing and it's a learning process. It
requires continuous reflections, dialogue, and call to actions. As
he talked about who is benefiting what. And you know, economic
gain and all of that. And I think that together we can strive for
a digital space that is inclusive and respects and empowers all
individuals, all communities regardless of their background,
regardless of their geographical location. We have to work
together in order to create a future where the ipt net truly
becomes a force of equality. Justice and liberation.
Thank you. And that is it for this session.
Thank you all.
-END-
We have another session in -- I think --
>> ANANYA SINGH: Take it from there. We have another session
ap happening at 1730. Standard time. I hope you will stay with
us. If you want to grab something quickly to drink or eat
meanwhile and if you're going outside I would request you to bring
in your colleagues to join us for the next session. Thank you very
much for attending.