Winning the battle against online hate speech in Myanmar | Min Thu Aung | TEDxUM1Yangon

We are now living in world of increasing connectivity. This connectivity brings opportunities as well as challenges. One of the overarching challenges is our safety during online activities. Nowadays, we are seeing increasing hate speech, fake news, and other harmful materials on the internet. In this talk, Min Thu Aung analyzes the issue of online safety in the context of Myanmar and explores the ways in which we can tackle the problem. Min Thu Aung leads the Business Sustainability and Special Projects team at Telenor in Myanmar. Since 2016, he has been leading a program that has educated over 300,000 youth in Myanmar in being responsible digital citizens on social media, with a focus on tackling the pervasive issue of hate speech. Prior to this, Min led a microfinance institution in Myanmar focusing on financial inclusion and was a management consultant in the UK before his return to Myanmar in 2013.

As a son of two expatriate Myanmar doctors, he was educated in the culturally diverse environments of Malaysia and the UK, where he personally witnessed how diversity can drive sustainable, socioeconomic growth and social harmony. He believes that empowering future generations of Myanmar’s youth to appreciate diversity as responsible digital citizens and to refrain from hate speech on Myanmar’s social media platforms will aid Myanmar in its ongoing reintegration into the global community following decades of isolation.

This talk was given at a TEDx event using the TED conference format but independently organized by a local community. Learn more at https://www.ted.com/tedx

OK, (Burmese) Hello.

My name is Min Thu Aung,

and I'm trying to find a solution

to solve the problem
of online hate speech in Myanmar.

So you might ask,
"Why is this so important?"

After all, hate speech
is nothing new, right?

Countries all over the world
have experienced hate speech in some way.

And as a country we have
a lot of problems to solve.

I'll give you one. Healthcare.

We need to improve our healthcare,

we need to improve our education
our roads, electricity maybe -

a lot of problems to solve.

Why hate speech? And why now?
Well, here's why.

As we know well,

back in August 2017, there
was the Rohingya crisis of course.

The UN fact-finding mission

concluded that online
hate speech on Facebook

was one of the main reasons
for this crisis.

This is hate speech
that divided communities

and caused a refugee crisis.

This is hate speech that -

really damaged our country's
international reputation,

and also reduced

the investment friendliness
of the country from a Western perspective.

It also is hate speech
that one day will affect all of us,

and, therefore, I believe
that it's an important problem to solve.

Here's another reason:

Just last month,

an LGBT librarian
at a private university in Yangon

was forced to confess his status
in a case of workplace harassment.

This harassment continued online
where he experienced cyberbullying

and also hate speech
because of his sexual orientation

and, tragically, he took his own life.

It's a big problem.

Another reason.

Another reason is that the elections
are coming up in 2020.

Political parties now are already
taking their positions,

and just like the US
election in 2016 also,

we expect that hate speech
could become, and fake news,

could be a feature
of the elections in 2020.

So that's why I'm convinced
that online hate speech,

in any democracy,

is an important issue to solve,
and Myanmar is no exception.

It applies to Myanmar as well.

I guess we first need
to look at what is hate speech?

What is the problem we're trying to solve?

What is free speech?

Where does free speech end,
and where does hate speech begin?

And for that, we can rely on -

international definitions
of hate speech and free speech.

So article 19 of the Universal
Declaration of Human Rights

and also the International Covenant
on Civil and Political Rights, ICCPR,

says that free speech is the right
to hold any opinion you want

and also to seek and receive
information in any way, right?

That's quite clear. You can say
anything you want. That's free speech.

But article 20 of the ICCPR
goes one step further.

It tells you where
your free speech could be limited.

If your free speech includes
any propaganda for war,

any advocacy of national, racial,
or religious hatred

which incites discrimination,
incites hostility or violence,

then your free speech can be restricted.

That's pretty clear.

But what does that mean
in normal language? That means ...

Well, in private
you can say anything you want,

even if it's very controversial,

but if you use that same speech in public

to incite any violence, any hostility
against specific groups of people,

then that can be considered hate speech,
and that can be offline or online.

It's the same.

Now, how big is the problem, right?

I think it's first important to note

that social media and the internet
is not all about hate speech, right?

If anything, the UN in 2016 passed
a resolution that says that the internet

is actually required or actually allows
the enjoyment of people's human rights.

It's very important, and some countries
have also even made internet access

a constitutional right, right?

So it's not the internet and social media.

It's not all about hate speech.

However, in 2014, when most
of the Myanmar population

got access to the internet,

they also got access to a tool
which allowed them

to access and express their views
to millions of people

for the first time in their lives.

Behind the veil of anonymity,

users on Facebook realized
that the speech they were using online

had consequences,

negative consequences for people offline.

And with 20 million users in Myanmar,

Facebook is effectively
the internet, right?

I think we all know that
in some way, shape, or form,

and therefore, for the rest of my talk
I'm going to focus on Facebook

as the main area where hate speech
exists online in Myanmar.

Now, concrete information or data
on hate speech is hard to find.

According to a survey done
by Myanmar mobile operator Telenor,

back in 2016,

the survey found that around 28%
of the people surveyed

knew someone that had been cyberbullied,
and hate speech could be a part of it.

But actually, anecdotally,
we know it's a big problem.

If you look at any Facebook page
of news outlets in the country,

you will find in the comments
following each article a lot of speech

which could be termed hate speech,
relating to religion, relating to race,

relating to other characteristics.

We also know, of course,
as I mentioned earlier,

that the UN fact-finding mission also said

that hate speech on Facebook
contributed to the Rohingya crisis.

So, in a nutshell,
it's a big problem, right?

But this is not a new problem.

Of course, countries in the West
have used social media and the internet

since the mid-nineties, right?

So how are they tackling the problem?

Well, I will show the way
that they are tackling this problem,

based on this framework
of the four solutions for hate speech.

First, let's look at
the state solution which is -

the governments regulate
and control speech online and offline.

One example of this
is actually in Germany.

They passed an act which
has been called the "Facebook Act,"

which requires Facebook
to remove any law-breaking content

within 24 hours of been told.

If they don't do this,

then they could be fined
up to 50 million euros.

And just in July this year,
Facebook was fined 2 million euros

for breaking some part of this law.

So this law is actually been described
as one of the most severe ways

that the government
and regulators around the world

have limited hate speech online.

And I know that France is also looking
at a similar law as well.

So civil societies, the NGOs
and rights groups in Germany

have actually criticized this law,

saying that it could actually
reduce free speech,

it could actually lead to censorship.

Why is that?

Well, Facebook can be fined
if there's any law-breaking content.

So Facebook could proactively
remove content that some people

might not even consider to be hate speech,
so that could lead to censorship.

Now, in a country like Germany,

which is a very advanced democracy,

if there's criticism
for this kind of a law,

in a newer democracy

where legislative, executive
and judicial capacity are still limited,

and freedom of the press
is still also limited,

this kind of a law is extremely risky.

But some countries
are looking at the solution.

Let's look at the social media solution.

And this is where platforms,
such as Facebook,

remove content proactively,

regardless of whether
there are laws or not.

Now, this solution is actually
quite active in a sense

that Facebook has a few thousand
content moderators around the world

that review content that has been flagged
by users as hate speech,

or as another kind
of non-compliant content.

Also in Myanmar,
I understand that Facebook

has a hundred or so content moderators
in Myanmar language,

and I also understand that Facebook

is in the longer term looking at using
artificial intelligence, or "AI,"

to help detect and review hate speech.

This can be really fast and scalable,
but under three conditions.

Number one: that these content moderators
actually can distinguish

between free speech
and hate speech without bias.

In an unbiased fair way. Number one.

Number two: the way
that hate speech is expressed

from time to time actually changes, right?

The way it's expressed today
and expressed next week may be different.

So it's important that content moderators

actually keep up to date with the way
that hate speech is expressed online.

That's the second condition.

And the third one is that Facebook
is investing in AI.

Well, it's important that this AI
is actually trained well

by the decisions made by users
and also by content moderators,

so that the AI can make its own decisions
accurately at a later point in time.

Of course, this solution
can be quite expensive for Facebook.

Lots of content moderators,
lots of new technology, lots of AI -

but I don't think Facebook, frankly,
has a choice but to invest in this.

They need to invest in this,

so it's not seen as a place
for hate speech and fake news

online in the future.

So we may have heard
of the Cambridge Analytica scandal

on Facebook.

I believe that if this hate speech issue
is not solved on Facebook,

it could become
the next Cambridge Analytica,

which could potentially be detrimental
for Facebook if they don't invest.

Now, looking at a third way,
the self solution,

which is where each of us actually
individually take responsibility

for what we post, what we share,
what we like online.

And there are three ways this can happen.

Number one is through formal education.

So online curriculum,
online training, online safety

being part of the national curriculum.

That's the first way.

The second way is through
informal training.

So for people outside of school,

they can be trained
by mass awareness programs,

by government on TV,
on radio, in newspapers.

That's another way.

And the third way is that we could train
people on Facebook itself.

Now, I have to say that Myanmar

is actually making
some progress in this regard,

so the Myanmar mobile operator
Telenor, for example,

has actually trained 300,000
students on hate speech

and online safety in the last
two-and-a-half years,

and this has been done collaborating
with the Ministry of Education,

with civil society group MIDO

and with training provider KMD as well.

All four partnering together.

Of course, we also know that online safety
is in the national curriculum in the UK,

in Germany, in Singapore,
and Sweden for example.

So this is a solution that could work,
long-term is probably required,

but because it requires behavioral change,

we have to change the way
we think, the way we act.

That takes time, and therefore,
it cannot be acting alone.

And last but not least,
the civil society solution:

so this is where NGOs work together
to actually determine what hate speech is,

and in Myanmar, this has been quite
an important part of reducing hate speech.

So there have been a number
of campaigns done by civil society.

So the Panzagar campaign by MIDO,
for example, is one example.

The Myanmar Institute
for Peace and Security

has an online hate speech tracker,

and there's been a lot of media literacy
and fact-checking campaigns

organized by civil society as well.

Civil society are very active
and they will remain active

I believe going forward.

Oops!

So what's the answer for us in Myanmar?

How can we actually apply
the learning from around the world

and create our own solution?

Well, let's use the same framework again.

And as I mentioned, the state solution -

so having laws to regulate hate speech,

maybe it's a little early
for that in Myanmar.

But I do think though that Facebook
could play a much more prominent role.

They need to hire more content moderators
familiar with Myanmar language

and also even the ethnic languages.

The second thing is we need
better content moderators:

content moderators
that can keep up-to-date

with the way that hate speech
is expressed online.

Content moderators
that can distinguish clearly

between free speech and hate speech,

and content moderators whose decisions
are reviewed regularly,

and as used as a way to train future
content moderators as well.

I think that Facebook cannot act alone.

They need to partner
closely with civil society,

and civil society can help Facebook
spot new trends in hate speech

and also help train and support
the content moderators,

so that they work hand in hand.

So that's in the short term.

In the medium term,
I mentioned earlier that I believe AI

artificial intelligence,

is a very important part
in reducing hate speech.

So with more content moderators,
collaboration with civil society,

I believe that these -

decisions made by content moderators

should be used
to train the AI of Facebook,

such that posts can be flagged, reviewed,
and maybe even taken down one day,

consistently in line
with international standards

of freedom of expression,

so that we're not relying
on human moderators only.

Because human moderators to moderate -

well, 100 to even 500 content moderators
to moderate 20 million people's speech,

it's a bit difficult.

We need a technical solution as well.

In the long term though,

I think hate speech
can really only be reduced

if we change the way we think as a people.

And that can be done in two ways I think:

First, I think there should
be mandatory online safety training,

including on hate speech
on Facebook itself.

New users should be required
to actually go to an online safety course

before they use Facebook.

Existing users, like probably you and me,

should take annual recertification
in Facebook's Community Guidelines,

which include online safety
and hate speech.

And until we complete
that recertification,

the account should be locked.

We should be forced to go through
this training every year by Facebook, OK?

Second thing is I believe it's important

to have an online safety course
in our national curriculum.

With that, I'm pleased to announce
that on 12th of July, just last week,

Telenor Myanmar and international NGO
Plan International signed an agreement

to explore including online safety
in the national curriculum,

together with the Ministry of Education.

We're working on this.

And if it works well, we will be educating

millions of our children every year

on how to be responsible
digital citizens online,

avoiding hate speech, and using
social media in a positive way.

So I believe that -

platforms, Facebook
and its content moderators,

its AI, civil society, educators,
all working together

can significantly reduce hate speech,
but what about you?

What can you do?

How can you be a part of the solution?

Or have you been a victim
of hate speech yourself?

Well, here are four tips to reduce
hate speech and what you can do.

The first one's obvious, right?

Avoid hate speech yourself.
That's a given.

Secondly, don't share or like
any post or comments

that contain hate speech.

Sometimes hate speech
can appear as fake news,

as jokes on Facebook as well.

So be careful not to share, not to like,
and not to comment on these posts either.

You can also report any instance
of hate speech directly to Facebook,

so that their content moderators
can review and quickly take down

any hate speech quickly.

And last but not least,
if you have been a victim yourself,

you're able to now report
to the cybercrime police in Naypyidaw.

They now have about 100 officers
that are looking at this issue.

Again, 100 officers for 20 million users.

It's still a bit small,
but it's a good start.

But you can also call on them,
or alternatively,

report the issue to your local
police station as well,

for assistance if you
are a victim of hate speech.

Now, I think for this solution,

for hate speech to be reduced
in Myanmar long-term,

we all need to be part of this solution,
and that starts now.

Thank you so much everyone.

(Applause)