#rC3 - The EU Digital Services Act package

rC3 preroll music

Herald: And the the talk that is about to
begin now is by Christoph Schmon from EFF

and Eliska Pirkova from Access Now and
they will be talking about this, probably

maybe the biggest legislative initiative
since the GDPR in Brussels. It's called

the Digital Services Act. And onto you
two. You're muted.

Eliska Perkova: I realized. Hello!
First of all, thank you very much for

having us today. It's a great pleasure to
be the part of this event. I think for

both of us, it's the first time we are
actually joining this year. So it's

fantastic to be the part of this great
community. And today we are going to talk

about the legislative proposal, the EU
proposal which causes a lot of noise all

around Europe, but not only in Europe, but
also beyond. And that's the Digital

Services Act legislative package that
today we already know that this

legislative package actually consists of
two acts: Digital Services Act and a

Digital Market Act. And both of them will
significantly change the regulation of

online platforms with a specific focus on
very large online platforms, also often

referred to as gatekeepers. So those who
actually hold a lot of economic dominance,

but also a lot of influence and control
over users rights and the public

discourse. So I'm going to start with
giving you a quick introduction into

what's the fuss about, what is actually
the DSA about, why we are also interested

in it and why we keep talking about it,
and why this legislation will keep us

preoccupied for the years to come. DSA was
already announced two years ago as a part

of European Union digital strategy, and it
was appointed as one of those actions that

the digital strategy will be actually
consisting of. And it was the promise that

the European Commission gave us already at
that time to create the systemic

regulation of online platforms that
actually places hopefully the users and

their rights into the center of this
upcoming legislation. So the promise

behind DSA is that these ad-based Internet
bill, and I'm speaking now about Ad-Tech

and Online-Targeting. The
Internet, as we actually knew, will be

actually replaced by something that puts
user and users controls, and users right

as a priority. So both of these
legislations implemented and drafted

right, should be actually achieving that
goal in the future. Now, previously,

before DSA actually was drafted, there was
some so-called e-Commerce Directive in

place that actually established the basic
principles, especially in the field of

content governance. I won't go into
details on that because I don't want to

make this too legalistic. But ultimately,
DSA legislation is supposed to not

completely replace but build up on the top
of this legislation that actually created

the ground and the main legal regime for
almost 20 years in Europe to regulate the

user generated content. So, DSA and DMA,
as the legislation will seek to harmonize

the rules addressing the problem, such as
online hate speech, disinformation. But it

also puts emphasis finally on increased
meaningful transparency in online

advertising, the way how the content is
actually being distributed across

platforms and also will develop a specific
enforcement mechanism that will be

actually looking into it. Now, before I
will actually go into the details on DSA

and why DSA matters, and do we actually
need such a big new legislative reform

that is coming from the European
Commission? I want to just unpack it for

you a little bit what this legislative
package actually consists of. So, as I

already mentioned, two regulations... the
regulation, the strongest legal instrument

European Commission actually has in its
hands, which is supposed to achieve the

highest level of harmonization across the
member states. And we all can imagine how

difficult that will be to achieve,
especially in the realm of freedom of

expression and particular categories of
user generated content, which is so deeply

complex dependance. All of those
related to content moderation and content

curation will be mainly in the realm of
Digital Services Act. And then the second

regulation, the Digital Market Act, will
be specifically looking at the dominance

of online platforms, economic dominance,
competitive environment for smaller

players, the fairness in the competition.
And it will also establish the list of

do's and don'ts for gatekeeper's
platforms. So something that... is so the

platforms that actually hold a relevant
dominance and now based on these new

proposals, we know that these platforms
are mainly called as very large online

platforms. So this is exactly how the
legislation refers to gatekeepers now. And

now I think one more point that I want to
make is that the DSA and the DMA were

launched on the 15th of December 2020. So it
was literally a Christmas present given to

the digital rights community by the
European Commission, a long time

anticipated one. The work on DSA
started, however much earlier. Electronic

Frontiers Foundations, as much as Access
Now, together with EDRi, were working very

hard to come up with the priorities and
recommendations what we would like to see

within these legislations to be
enshrined, because from the beginning we

understood the importance and the far
reaching consequences this legislation

will have not only inside of the European
Union, but also beyond. And that brings me

to the final introductory point that I
want to make before I will hand over to

Chris, which is why and do we actually
need DSA? We strongly believe that there

is a big justification and good reason to
actually establish a systemic regulation

of online platforms in order to secure
users fundamental rights, empower them,

and also to protect our democratic
discourse. And this is due to the fact

that for many years we're witnessing this
phase: quite bad regulatory practices in

platform governance that are coming from
member states and Chris will provide for a

very concrete example in that regard, but
also coming from the European Commission

itself, mainly the proposed online service
content regulation, for instance, or we

all remember the story of copyright that
Chris will discuss a little bit further.

We saw how not only the member states, but
also European Commission or European

Union, in order to actually establish some
order in the digital space, they started

pushing the state's obligation and
especially states positive obligation to

protect the users' human rights in the
hands of online private platforms that

started replacing state actors and public
authorities within the online space. They

started assessing the content, the
legality of the content, deciding under a

very short time frames what should stay
online and what should go offline with no

public scrutiny or transparency about
practices that they kept deploying. And

they keep deploying to this day. Of
course, platforms under the threat of

legal liability often had to rely and
still have to rely on the content

recognition technologies for removing user
generated content. A typical example could

be also Avia law, which will be still
mentioned today during the presentation.

And the typical time frames are usually
those that extend from one hour to 24

hours, which is extremely short,
especially if any users would like to

appeal such a decision or seek an
effective remedy. At the same time, due to

the lack of harmonization and lack of
proper set of responsibilities that should

lie in the hands of these online
platforms. There was a lack of legal

certainty which would only reinforce the
vicious circle of removing more and more

of online content in order to escape any
possible liability. And at the end, to

this day, due to the lack of transparency,
we lack any evidence or research based

policy making, because platforms do not
want to share or inform the public

authorities what they actually do with the
content, how they moderate and those

transparency information that we receive
within their transparency reports are

usually quantity oriented instead of
quality. So they focus on how much content

is actually being removed and how fast,
which is not enough in order to create

laws that can actually provide any more
sustainable solutions. And ultimately, as

we all agree, the core issue doesn't lie
that much with how the content is being

moderated, but how content is being
distributed across platforms within the

core of their business models that
actually stands on a attention economy and

on the way, how sensational content is
often being amplified in order to actually

prolong that attention span of users that
visit platforms on regular basis. And I'm

packing quite a few issues here. And this
supposed to be just like a quick

introductory remark. I will now hand over
to Chris that will elaborate on all these

points a little bit further, and then we
take a look and unpack a few quite

important parts of the essay that we feel
should be prioritized in this debate at

the moment. Chris, over to you.
Christoph Schmon: Hi everybody. I'm

quite sure that many of you have noticed
that there's a growing appetite from the

side of the European Union to regulate the
Internet by using online platforms, as the

helping hands to monitor and censor what
users can say/share/do online. As you

see on the slide, the first highlight of
this growing appetite was corporate upload

filters, which are supposed to stop
copyright protected content online.

Thousands of people, old and young, went
on the streets to demonstrate for free

Internet, to demonstrate against technical
measures to turn the Internet into some

sort of censorship machine. We've made a
point then, and we continue making the

point now that upload filters are prone to
error, that upload filters cannot

understand context, that they are
unaffordable by all but the largest tech

companies, which happen to be all based in
the United States. But as you well know,

policymakers would not listen and
Article 13 of the copyright directive was

adopted by a small margin in the European
Parliament, also because some members of

European Parliament had troubles to press
the right buttons, but I think it's

important for us to understand that the
fight is far from over. The member states

must now implement the directive in the
way that is not at odds with fundamental

rights. And we argue that mandated
automated removal technologies are always

in conflict with fundamental rights. And
this includes data protection rights. It

is a data protection right not to be made
subject to Automated Decision-Making

online, if it involves your personal data
and if such decision making has a negative

effect. So we believe, all those legal
arguments aside, I think the most worrying

experience with upload filters is that it
has spillover effects to other

initiatives. Sure, if it works for
copyright protected companies, it may well

work for other types of content, right?
Except that it doesn't. Many considered

now a clever idea that web forums should
proactively monitor and check all sorts of

user content. May this be communication,
pictures or videos, and they should use

filters to take it down or the filters to
prevent the re-upload of such content. An

example of such spillover effect that Eliska
had mentioned is the draft regulation of

terrorist related content. It took a huge
joint effort of civil society groups and

some members of Parliament to reject the
worst of all text. We had recent

negotiations going on and at least we
managed to get out the requirement to use

uploads filters, but still a twenty four
hours removal obligation that may nudge

platforms to employ those filters
nevertheless. And we see that particularly

in national states, they are very fond of
the idea that platforms rather than

judges should be the new law
enforcers. There are now several states in

the European Union that have adopted laws
that would either oblige or nudge

platforms to monitor users speech
online. First up was the German NetzDG,

which set out systematic duties for
platforms. Then we had the French law

Avia, which copy-pasted the NetzDG and
made it worse. And last we have the

Austrian Hate Speech bill, which is a mix
of both the German and the French

proposal. They all go much beyond
copyright content, but focus on hate

speech and all sorts of content that may
be considered problematic and in those

respective countries, not necessarily in
other countries. And this brings me to the

next problem. How do we deal with content
that is illegal in one country, but legal

in another? A recent Court of Justice
ruling had confirmed that a court of a

small state like Austria can order
platforms not only to take down defamatory

content globally, but also to take down
identical or equivalent material using

automated technologies. For us this is a
terrible outcome, that will lead to a race

to the bottom, where the countries with
the least freedom of speech friendly laws

can superimpose their laws on every other
state in the world. We really believe that

all this nonsense has to stop. It's time
to acknowledge that the Internet is a

global space, a place of exchange of
creativity and the place where civil

liberties are suppose to exist. So we are
fighting now against all those national

initiatives. We had the great first
victory, when we have to bring down the

French law Avia - the Avia bill that had
imposed the duty for platforms to check

and remove potentially illegal content
within 24 hours. Before the Conseil

constitutionnel, the French Supreme Court,
we had argued that this would push

platforms to constantly check what users
post. And if platforms face high

fines... of course, they would be rather
motivated to block as much context of the

content as possible. We've made a point
that this would be against the Charter of

Fundamental Rights, including freedom of
information and freedom of expression. And

it was a great victory for us that the
French Supreme Court has struck down the

French Avia bill and followed our
argument, as you see on the slide. We also

see now that there's a push back at least
against the update of the German NetzDG,

which would have provided new access
rights for law enforcement authorities.

This and other provisions are considered
unconstitutional. And as far as I

understand and, perhaps listeners can
correct me, the German president has

refused to sign the bill and the Austrian
bill goes the similar path way - got

recently a red light from Brussels. The
commission considers it in conflict with

EU principles. Also, thanks to joint
effort by epicenter.works. And this shows

that something positive is going on, it's
a positive development, the pushback

against automated filter technologies.
But it's important to understand that

those national initiatives are not just
purely national attempts to regulate hate

speech. It's an attempt of an EU member
state to make their own bills, as badly as

they are, some sort of a prototype for EU-
wide legislation, a prototype for the

Digital Services Act. And as you know,
national member states have a say in EU

lawmaking: their voices are represented in
the council of the EU and the European

Commission, which will be disincentivized
to propose anything that would be voted

down by the council. I think that's a nice
takeaway from today, that lawmaking in

national member states is not an isolated
event. It's always political, it's always

Netzpolitik. The good news is that as
far as the European Commission proposal

for a Digital Services Act is concerned,
that it has not followed the footsteps of

those bad, badly designed and misguided
bills. It has respected our input, the

input from Access Now, from EFF, from the
EDRi network, from academics and many

others, that some key principles should
not be removed, like that liability for

speech should rest with the speaker. The
DSA, it's also a red light to channel

monitoring of users content. And there are
no sure badlines in there to remove

content that might be illegal. Instead,
the commission gives more slack to

platforms to take down posts in good faith,
which we call the EU style

Good Samaritan clause. Looking
through the global lenses of law making

it's very fascinating to see that while
the United States is flirting with the

idea to move away from the Good Samaritan
principle in Section 230 of the

Communications Decency Act, so the idea is
that platforms can voluntarily remove

content without being held liable for it,
the European Union flirts with the idea to

introduce it, to give more options to
platforms to act. That being said, the

major differences between the US and the
EU is that in Europe, platforms could be

held liable the moment they become aware
of the illegality of content. That's an

issue because the Digital Services Act has
now introduced a relatively sophisticated

system for user notification, complaint
mechanism, dispute resolution options,

which all leads to such awareness about
illegality or could lead to such

awareness. It's not quite clear for us how
platforms will make use of voluntary

measures to remove content. That being
said, we think that the commission's

proposal could have been much worse. And
the Parliament's reports on Digital

Services Act have demonstrated that the
new parliament is a bit better than the

old one. They have a lot of respect for
fundamental rights. So many members of

parliament that are quite fond of the idea
to protect civil liberties online. But we

know that this was only the start and we
know that we need another joint effort to

ensure that users are not monitored, they
are not at the mercy of algorithmic decision

making. And I think Eliska is now going to
explain a bit more about all this.

Eliska: Thank you. Thank you very much,
Chris. So, we can actually move now

further and unpack a few relevant
provisions for everything that has already

been mentioned, mainly in the realm of
content moderation and content creation,

which ultimately lies in the core of
Digital Services Act. And maybe not to

make it all so abstract... I also have the
printed version of the law here with me.

And if you look at it, it's quite an
impressive piece of work that the European

Commission did there. And I have to say
that even though it's a great start, it

still contains a lot of imperfections. And
I will try to summarize those now for you,

especially in the light of our end positioning
and as I mean civil societies in general,

because I believe that on all those
points, we have a pretty solid agreement

among each other. And what we were
actually hoping that Digital Services Act

will do, what it actually does and where
we see that we will need to actually

continue working very closely, especially
with the members of the European

Parliament in the future, once the draft
will actually enter the European

Parliament, which should happen relatively
soon. So as it already, as I already

mentioned at the beginning, quite briefly,
is how actually DSA distinguishes between

online platforms, which are defined within
the scope of the law and between very

large online platforms, which is exactly
that scope where all large online

gatekeepers fall into. DSA specificly then
distinguishes between obligations or

responsibilities of these actors, some
assigning to all of them, including online

platforms and some being extended
specifically due to the dominance and

power of these online gatekeepers hold.
This is mainly then the case when we

discuss the requirements for
transparency. There is a set of

requirements for transparency that apply
to online platforms, but then there is

still specific additional set of
transparency requirements for larger

online platforms. What DSA does especially
and this is the bed which is extremely

relevant for the content moderation
practices - it attempts to establish a

harmonized model for notice and action
procedure for allegedly illegal content.

Whatever alert red lines before I go into
the details on this was that DSA will be

touching only, or trying to regulate only
allegedly illegal content and stay away

from vaguely defined content categories
such as potentially harmful, but legal

content. There are other ways how the
legislation can tackle this content,

mainly through the meaningful transparency
requirements, accountability, tackling

issues within the open content recommender
systems and algorithmic curation. But we

didn't want the specific category of the
content to be included within the scope of

DSA. This is due to the fact that vaguely
defined terms that find their way into

legislation always lead to human rights
abuse in the future. I could give you

examples from Europe, such as the concept
of online harms within the UK, but also as

a global organizations. Both of us, we
actually often see how weak terminology

can quickly lead to even over
criminalization of speech or suppressing

the decent. Now, if we go back to
harmonize notice and action procedure,

what that actually means in practice, as
Christoph already mentioned, Europe has

so-called conditional model of
intermediate reliability, which is being

provided already and established by the
initial legal regime, which is the

e-Commerce Directive under Article 14 of
the e-Commerce Directive, which actually

states that unless the platform holds the
actual knowledge and according to the

wording of DSA now it's the actual
knowledge or awareness about the presence

of illegal content on their platform, they
cannot be held liable for such a content.

Now, we were asking for a harmonized
procedure regarding notice and action across

the EU for a while, precisely because we
wanted to see reinforced legal certainty.

Lack of legal certainty often translated
into overremoval of even legitimate

content from the platforms with no public
scrutiny. DSA does a good job, it's a good

starting point that actually tries to
attempt, to establish such a harmonized

procedure, but it's still lacking behind
on many aspects that we consider important

in order to strengthen protection of
fundamental rights of users. One of them

is, for instance, that harmonized notice
and action procedure, as envisioned by

DSA, is not specifically tailored to
different types of categories of user

generated content. And as we know, there
were some or many categories of content

that are deeply context dependent, linked
to the historical and sociopolitical

context of member state in question. And
due to their reliance on the automated

measures, that usually context blind, we
are worried that if notice and action

doesn't reflect this aspect in any further
ways we will end up again with over

removals of the content. What is probably
another huge issue that we are currently

lacking in the draft, even though DSA is
trying to create a proper appeal and

enforcement mechanisms and also appeal
mechanisms for users and different

alternative dispute settlement of the law
draft currently contains, there is no

possibility for content providers, for the
user that uploads the filter.. Sorry.

laughs That was a nice Freudian slip
there. For

a user that actually uploaded content to
appeal to directly actually use the

counter notification about that notified
content that belongs to that user. Nor

platforms are obliged to actually send the
notification to a user prior to any action

that is being taken against that
particular notified content. These are for

us a procedural safeguards for fairness
that users should have, and currently they

are not being reflected in the draft.
However, this is a good start and it's

something that we were pushing for. But I
think there are many more aspects that

these notice and action procedures will
need to contain in order to truly put

users at first. Now the notice and action
procedure is mainly focusing on the

illegal content. But there are ways in the
draft where potentially harmful content,

which is still legal - so the content that
actually violates the terms of service of

platforms is being mentioned throughout
the draft. So for us, it's now at the

moment exactly clear how that will work in
practice. So that's why we often use this

phrase that is also put on the slide: good
intention with imperfect solutions.

However, I want to emphasize again that
this is just the beginning and we will

still have time and space to work very
hard on this. Another kind of novel aspect

that DSA actually brings about is already
mentioned Good Samaritan Clause, and I

tend to call it the EU model or EU version
of Good Samaritan Clause. Good Samaritan

Clause originates in Section 230 of
Communication Decency Act, as Cristoph

already mentioned. But within the European
realm, it goes hand in hand with this

conditional model of liability which is
being preserved within the DSA legal

draft. That was also one of our main ask
to preserve this conditional model of

liability and it's great that this time
European Commission really listened. Why

we consider the Good Samaritan Clause
being important? In the past, when such a

security wasn't enshrined in the law, but
it was just somehow vaguely promised to

the commission that if the platform will
proactively deploy measures to fight

against the spread of illegal content,
they won't be held liable without

acknowledging that through such a use of
so-called proactive measures, the platform

could in theory gain the actual knowledge
about the existence of such a type of

content on its platform, which would
immediately trigger legal liability. This

threat of liability often pushed platforms
to the corner so they would rather remove

the content very quickly then to face more
serious consequences later on. That's why

we see the importance within the Good
Samaritan Clause or the European model of

Good Samaritan Clause, and we are glad
that it's currently being part of the

draft. One of the biggest downfalls or one
of the biggest disappointments when DSA

finally came out on the 15th of December
for us was to see that it's still online

platforms that will remain in charge when
it comes to assessing the legality of the

content and deciding what content should
be actually restricted and removed from a

platform and what should be available. We
often emphasize that it's very important

that the legality of the content is being
assessed by the independent judicial

authorities as in line with the rule of
law principles. We also do understand that

such a solution creates a big burden on
the judicial structure of member states.

Many member states see that as a very
expensive solutions, they don't always

want to create a special network, of
courts, or e-courts or other forms of

judicial review of the illegal or
allegedly illegal content. But we still

wanted to see more public scrutiny,
because for us this is truly just the

reaffirmation of already existing status
quo, as at the moment and there are many

jurisdictions within the EU and in the EU
itself, it's still online platform that

will call the final shots. What, on the
other hand, is a positive outcome that

we were also hardly pushing for are the
requirements for meaningful transparency.

So to understand better what platforms
actually do with the individual pieces of

content that are being shared on these
platforms and how actually transparency

can then ultimately empower user. Now, I
want to emphasize this because this is

still ongoing debate and we will touch
upon those issues in a minute. But we

don't see transparency as a silver bullet
to the issues such as amplification of

potentially harmful content or in general
that transparency will be enough to

actually hold platforms accountable.
Absolutely not. It will never be enough,

but it's a precondition for us to actually
seek such solutions in the future. DSA

contains specific requirements for
transparency, as I already mentioned, a

set of requirements that will be
applicable largely to all online platforms

and then still specific set of
requirements on the top of it. That will

be applicable only to very large online
platforms so the online gatekeepers. We

appreciate the effort, we see that the
list is very promising, but we still think

it could be more ambitious. Both EFF and
Access Now put forward a specific set of

requirements for meaningful transparency
that are in our positions. And so did EDRi

and other civil society or digital rights
activists in this space. And final point

that I'm going to make is the so-called
Pandora box of online targeting and

recommender systems. Why do I refer to
this as to Pandora Box? When a European

Parliament published its initiative
reports on DSA, there are two reports, one

being tabled by JURI Committee and then
another one by ENCO, especially the JURI

report contained paragraph 17, which calls
out for a better regulation of online

targeting and online advertisements, and
specifically calling for a ban of online

targeting and including the Phase-Out that
will then lead to a ban. We supported this

paragraph, which at the end was voted for
and is the part of the final report.

Nevertheless, we also do understand that
these wording of the article has to be

more nuanced in the future. Before I go
into the details there, I just want to say

that this part has never made it to DSA.
So there is no ban on online targeting or

online advertisement of any sort, which to
us, to some extent, it was certainly

disappointing too, we specifically would
call for a much more stricter approach

when it comes to behavioral targeting as
well as crossside tracking of online

users, but unfortunately, and as we
eventually also heard from Commissioner

Vestager, that was simply lag of will or,
maybe, too much pressure from other

lobbies in Brussels. And this provision
never found its way to the final draft of

DSA. That's the current state of art, we
will see what we will manage to achieve

once the DSA will enter the European
Parliament. And finally, the law also

contains a specific provision on
recommender systems. So the way how the

content is being distributed across
platform and how the data of users are

being abused for such a distribution and
personalization of user generated content.

In both cases, whether it's online
targeting and recommender systems within

the DSA, DSA goes as far as the
transparency requirements, the

explainability, but it does very little
for returning that control and empowerment

back to the user. So whether user can
obtain or opt out from these algorithmic

curation models, how it can actually be
optimized if they decide to optimize it?

All of that is at the moment very much
left outside of the scope of DSA. And so

that's the issue of interoperability,
which is definitely one of the key

issues being currently discussed and made
kind of possible hopes in the future for

returning that control and empowerment
back to the user. And I keep repeating

this as a mantra, but it's truly the main
driving force behind all our initiatives

and the work we do in these fields. So the
user and their fundamental rights. And on

that note, I would like to hand over back
to Chris, who will explain the issue of

interoperability and how to actually
empower you as a user and to strengthen

the protection of fundamental rights
further. Chris, it's yours now. Christoph:

Thank you. I think we all know or feel
that the Internet has seen better times.

If you look back over the last 20 years,
we have seen that transformation was going

on from an open Internet towards a more
closed one - monopolization. Big platforms

have built entire ecosystems and it seems
that they alone decide who gets to use

them. Those platforms have strong network
effects that have pushed platforms or

those platforms into gatekeeper position
which made it so easy for them to avoid

any real competition. This is especially
true when we think of social media

platforms. This year we celebrate the 20th
birthday of the e-Commerce Directive that

Eliska mentioned. The Internet bill that
will now be replaced by the Digital

Services Act. We believe it's a very good
time now to think and make a choice:

should we give even more power to the big
platforms that have created a lot of the

mess in the first place; or should we give
the power to the users, give the power

back to the people? For us, the answer is
clear. Big tech companies already employ a

wide array of technical measures. They
monitor, they remove, they disrespect user

privacy and the idea to turn them into
the Internet Police, with a special

license of censoring the speech of users,
will only solidify their dominance. So we

wouldn't like that. What we like is to
put users in charge over their online

experience. Users should, if we had a say,
choose for themselves which kind of

content they can see, what services they
can use to talk to their friends and

families. And we believe it's perhaps time
to break up those silos, those big

platforms have become to end the dominance
over data. One element to achieve this

would be to tackle the targeted ads
industry, as Eliska mentioned it, perhaps

to give an actual right to users not to be
subject to targeted ads or to give more

choice to use to decide, which content
they would like to see or not to see. In

the Digital Services Act, the Commission
went for transparency when it comes to ads

and better option for users to decide on
the recommended content, which is a start,

we can work with that. Another important
element to achieve user autonomy over data

is interoperability. If the European Union
really wants to break the power of those

data driven platforms that monopolize the
Internet, it needs regulations that

enables users to be in control over the
data. We believe that users should be able

to access data, to download data, to move,
manipulate their data as they see fit. And

part of that control is to port data
from one place to another. But data

portability, which we have under the
GDPR is not good enough. And we

see from the GDPR that it's not
working in practice. Users should be able

to communicate with friends across
platform boundaries, to be able to follow

their favorite content across different
platforms without having to create several

accounts. But to put it in other terms, if
you upset with the absence of privacy on

Facebook or how the content is moderated
on Facebook, you should be able to just

take your data with you using portability
options and move to an alternative

platforms, that is a better fit and this
without losing touch with your friends who

stay behind, who have not left the
incumbent big platform. So what we did for

Digital Services Act is to argue for
mandatory interoperability options that

would force Facebook to maintain APIs that
let users on other platforms exchange

messages and content with Facebook users.
However, if you look in the DSA, we see

that the commission completely missed the
mark on interoperability, which is

supposed to be dealt with by related legal
act, now it gets complicated. It's the

Digital Markets Act, the DMA, another
beautiful acronym. The Digital Markets Act

wants to tackle certain harmful business
practices by those gatekeeper platforms,

the very large tech companies that control
what is called core services. The core

service is a search engine, a social
networking service, a messaging service,

its operating systems and online
intermediation services. Like think of how

Amazon controls access to customers for
merchants that sell on its platforms or

how the Android and iPhone app stores as
chokepoints in delivering mobile software.

And many things we like in the new
proposal, the proposal of the Digital

Markets Act, for example there's a ban on
mixing data in there that you may wants to

ban gatekeeper's from mixing data from
data progress with the data they collect

on the customers. Another rule is to ban
cross tying - sort of practices that end

users must sign up for ancillary services.
So you should be able to use Android

without having to get a Google account for
example. You believe that this is all

good, but the DMA like the DSA is very
weak on interoperability. What it does is

to focus on real time data portability
instead. So instead of having

interoperable services, users will only be
able to send the data from one service to

another like from Facebook to Diaspora,
meaning that you would end up having two

accounts instead of one or to quote Cory
Doctorow who spoke yesterday already:

"Users would still be subject to the
sprawling garbage novela of abusive legalese

Facebook lovably calls its terms of
service." We believe that this is not

good enough. And the last slide, you see a
quote from the Margrethe Vestager who made

a very good statement last month, that we
need trustworthy services, fair use of

data and free speech and an interoperable
internet; we fully agree on that. And in

the next months and years, we will work on
this to actually happen. However, you can

imagine, it will not be easy. We already
see that European Union member states

follow the trend of platforms should
systematically check undesirable and

insightful content and share those data
with enforcement authorities, which is

even worse. We see an international trend
going on to move away from the immunity of

platform for use of content towards a more
active stance of those platforms. And we

see that recent terror attacks have fueled
ideas that monitoring is a good idea and

end to end encryption is a problem. So
whatever will be the result, you can bet

that European Union will want to make the
Digital Services Act and the Digital

Markets Act another export model. So this
time we want the numbers right in

parliament and the council, we want to
help members of parliament to press the

right buttons. And for all this we will
need your help, even if it means to learn

yet another acronym or several acronyms
after the GDPR. That's it from

our side - we are looking forward to the
discussion.Thank you.

Herald: OK, thank you Eliska and

Christoph. There are questions from the
internet and the first one is basically

we just have and as you mentioned in your
slides, Christoph, the copyright in the

digital single market with both
accountability and liability provisions,

you also briefly mentioned, I think even
the e-evidence proposal also. How do all

these proposals relate to each other? And
especially for a layperson, that is not

into all the Brussels jargon.
Christoph: I think Eliska, you raised

your hand, don't you?
Eliska: ...more or less unintentionally,

but yeah, kind of that. I can start and
then let you Christoph to step in. Yeah,

that's a very, very good question. And
this is specifically due to the fact that

when you mention especially online
terrorist content regulation, but also

recently proposed interim regulation on
child sexual abuse, they.. all these - we

call them sectoral legislation, so kind of
a little bit of parting from this

horizontal approach, meaning an approach
that tackles all categories of illegal

content in one way, instead of going after
specific categories such as online

terrorist content in the separate ways. So
it's a little bit paradoxical saying what

is currently also happening at the EU
level, because on one hand, we were

promised this systemic regulation that
will once for all establish harmonized

approach to combating illegal content
online and at the same time, which is

specifically DSA, the Digital Services
Act, and at the same time we still see

European Commission allowing for these
fundamental rights harmful legislative

proposals happening in these specific
sectors such as proposed online

terrorist content regulation or other
legislative acts seeking to somehow

regulate specific categories of user
generated content. This is quite puzzling

for us as a digital rights activists too,
and very often, actually, so I would maybe

separate DSA from this for a moment and
say that all of these sectoral

legislations what they have in common is:
first of all, continuing these negative

legislative trends that we already
described and that we constantly observe

in practice, such as shifting more and
more responsibility on online platforms.

And at the same time, what is also very
interesting, what they have in common is

the legal basis that they stand on, and
that's the legal basis that is rather

connected to the cooperation within the
digital single market, even though they

seek to tackle a very particular type of
category of content category, which is

manifestly illegal. So logically, if they
should have that appropriate legal ground,

it should be something more close to
police and judicial cooperation, which we

don't see happening in practice,
specifically because there is this idea

that platforms are the best suited to
decide how the illegal content will be

tackled in the online space. They can be
the fastest, they can be the most

effective. So they should actually have
that main decision making powers and

forced into taking those responsibilities
which have ever ultimately, according to

the rule of law principle, should and have
to be in the hands of the state and public

authorities, preferably judicial
authorities. So I would say they are all

bad news for fundamental rights protection
of online users, civil rights

organizations, all of us that are on this
call today. We're fighting very hard also

against the online service content
regulation. There was a lot of damage

control done, especially with the first
report that was tabled by the European

Parliament and also now during the last
trialogue since the negotiations seems to

be concluded and the outcome is not great.
It's far from ideal. And I'm worried that

with other sectoral legislative attempts
coming from the European Commission, we

might see the same outcome. It will be
very interesting to see how that will

actually then play together with the
Digital Services Act, which is trying to

do the exact opposite to actually fix this
negative legislative efforts that we see

at the EU level with these sectoral
legislation, but also with the member

states at the national level. I could also
mention the European Commission reaction

to some national legislative proposals.
But Christoph, I would leave that to you

and please step in.
Christoph: I think you explained it

perfectly, and the only thing I can
supplement here is that if you look at

this move from sectoral legislation,
asylum legislation to horizontal

legislation, now back to sectoral
legislation - it's a problem, it's a mess.

First, the two sides not very good
coordinated which brings troubles for

legal certainty. It makes it very
troublesome for platforms to follow up.

And it's problematic for us, for us in the
space. We are some sort of lobbyist as

well, just for public interest. But you
will have to have to deal with copyright,

with CSAM, with TERREG, with end to
end encryption, DSA, DMA and 15 other

parties to pop up content by content. It's
very hard to manage to have the capacity

ready to be early in the debate, and it's
so important to be early in the debate to

prevent that from happening. And I think
that's a huge challenge for us, to have

something for us to reflect to in the next
days. How can we join forces better in a

more systematic way in order to really
follow up on all those initiatives? That's

for me, a very problematic development.
Herald: So in summary it's a mess. So it

is related, but we can't explain how,
because it's such a mess. Fair enough.

I have another question for you, Eliska.
Someone was asking how the proposed

Good Samaritan clause works compared to..
how it currently works in Germany. But I

think it's a bit unreasonable to expect
everyone to know how it works in Germany.

I would rephrase it this as: how does this
proposed Good Samaritan clause work

compared to how it is now under the
e-Commerce Directive?

Eliska: Thank you very much. Yeah, so a
great question again, I think the first if

we put it into the context of the EU law
and apologies that I cannot really answer

how - you know, compare the German context
- I really don't dare to, I'm not a

German lawyer, so I wouldn't like to step
in those waters. But first of all, there

is no Good Samaritan clause per se within
the scope of e-Commerce Directive. It did

not really exist within the law. And I'm
using the pass sentence now because DSA is

trying to change that. So that level of
legal certainty was not really, really

there for the platforms. There was the
conditional model of the liability, which

is still preserved within the regulation.
But if you think of a Good Samaritan

clause as we know it from the section
230, or let's use that Good Samaritan

clause as an example, because also
e-Commerce Directive was actually drafted

as a response to Communication Decency Act
that was the legislation that puts things

into motion. So that's the first
ultimate point. I explain at the

beginning in my presentation what was then
happening in the space of combating

illegal content at the EU level, and
especially I would refer to the

communication that the European Commission
published, I think, back in 2018, where it

actually encouraged and called on online
platforms to proactively engage with

illegal content and use these proactive
measures to actually seek an adequate

response to illegal content. Now, to mix
that with this conditional model of

liability, which is of course defined by
the obtaining actual knowledge by the

platform that created a perfect storm that
I already explained. So the platforms knew

that they are kind of pushed by the
legislature to actually seek these active

responses to illegal content, often
deploying automated measures. But they

didn't have any legal certainty or
security on their side that if they do so,

they won't end up ultimately being held
legally liable and face legal consequences

as a result of obtaining actual knowledge
through those proactive measures that were

kind of the tool, how they could possibly
actually obtain that knowledge. Now, what

DSA does, it specifically actually simply
states and I think it's Article 6 in the

Digital Services Act, if I'm not mistaken,
and I can even open it, it specifically

basically says that platforms can use
these proactive measures or, you know,

continue using some tools that actually
seek to provide some responses to this

type of content without the fear of being
held liable. So it's it's an article which

has approximately, I think, two
paragraphs, but it's finally in the

legislation and that means that it will
help to reinforce the level of legal

certainty. I would also emphasize that
very often in Europe, when we discuss Good

Samaritan clause, and Good Samaritan is
actually very unfortunate term, because

it's very much connected to the American
legal tradition. But when it's being mixed

up with the conditional model of liability
and with the prohibition of general

monitoring, which is still upheld, these
are the main principles of the European

intermediary reliability law and the
regime that is applicable within the EU,

such a safeguard can be actually
beneficial and it won't lead hopefully to

these blanket immunity for online
platforms or to this idea that platforms

will be able to do whatever they want with
the illegal content without any public

scrutiny, because there are other
measures, safeguards and principles in

place as a part of conditional model of
liability that we have here in Europe. So

I'm sorry, maybe that was too complicated.
Legalistic explanation there. But this is

how these provisions should work in
practice. We, of course, have to wait for

the implementation of the law and see how
that will turn out. But the main purpose

is that this legal certainty that was
lacking until now can finally come to its

existence, which should help us to prevent
over removal of legitimate speech from

online platforms.
Herald: OK, thank you. I have two other

questions from the Internet about
interoperability, and I suppose I should

look at Christoph for them. The last one
I'm going to ask first is: would such

interoperability make it much more
difficult to combat harassment and

stalking on the Internet? How do you
police that kind of misbehavior if it's

across different platforms who are forced
to interoperate and also be conduits for

such bad behavior. And I'll come to the
earlier question if you've answered this

question Christoph.
Christoph: It's a pretty good question.

First, to understand our vision on
interoperability is to understand that we

would like to have it between platforms
that empower large platforms and the right

of smaller platforms, actually, to make
use of interoperability. So it should not

be among the big platforms. So small
platforms should be able to connect to the

big platforms. And second, we believe it
will help and not make it worse because we

have now a problem of hate speech, we have
now a problem of a lack of privacy, we

have now a problem of the attention
industry that works with, you know,

certain pictures put in certain frames to
trigger the attention of users, because users

don't have a choice of the content
moderation practices, users don't have a

choice to see which kind of content should
be shown. And users don't have options to

regulate the privacy. The idea of more
competitors would be exactly that I can

move to a space, where I'm not harassed
and not be made subject to certain content

that hurt my feelings. Right. And that
moment I get control. I can choose a

provider that gives me those options and
we would like even to go a step further.

Back end interoperability was a start. We
believe if users want to, they should be

able to delegate a third party company or
piece of a third party software to

interact with the platform on their
behalf. So users would have the option to

see a news feed in different order,
calibrate their own filters on

misinformation. So in this sense,
interoperability can be a great tool,

actually, to tackle hate speech and to
sort of negative developments. Of course,

there is a risk to it. I think the risk
comes rather from the data industry side

again, that we need to take care not to
place one or another data selling

industry on the one that we already face.
But for this, we have options as well to

avoid that from happening. But to answer
the question, we believe interoperability

is a tool actually to escape from the
negative developments you had mentioned.

Herald: Critical counter question for me
then, aren't you actually advocating for

just roll your own recommendation engines
to be able to do so? Can't you achieve

that without interoperability?
Christoph: Sure. Recounter question: Do

you think an average user can accomplish
that quite easily? You know, like when we

look at the Internet through the lenses of
market competition then we see that it is

the dominance of platforms over data that
have created those spaces, those developed

gardens where users have the feeling they are
trapped and cannot escape from. And there

are so many alternative options that can
not get off the ground because users feel

trapped, don't want to leave their friends
behind and don't have options, actually to

have a better moderation system. Of course,
you can be creative and, you know, use

plugins and whatever you see fit, but you
need to stay within the platform barriers.

But we would like to enable users to actually
leave developed garden, go to another

place, but still stay in touch with
friends who have made the choice to remain

there. And I think that's perhaps the
difference to what you had in mind.

Herald: I have a follow up question. Well,
another question from the Internet,

regardless of interoperability, and that
is, historically speaking, as soon as the

big players get involved in certain
standards, they tend to also shape policy

by being involved in that. How would that
be different in the case of

interoperability and specifically
mentioned by the person who ask the

question. That Mastodon probably
flourishes because nobody else was

involved in setting that standard.
Christoph: Ah it's an excellent question.

And we struggled with the question of
standards ourselves in our policy paper,

which is our recommendations for European
Union to enact certain provisions in the

new Digital Services Act. We abstain from
asking to establish new standards like API

standards. We believe it's a bad idea to
regulate technology like that. What we

want to do is that big platforms just
offer interoperability however they see fit.

We don't want to have a standard that can
be either again monopolized or lobbied by

the big platforms. Because then we end up
with the standards we already see

which we don't like. But it's a good
question. And what we did is with our

policy principles on interoperability to
give kind of a food for thought, how we

believe the end version should look like,
but the many questions that remain and we

don't know exactly the way how to go
there.

Herald: Yeah, I'm sorry of sticking to the
topic of interoperability, because most

questions are actually about that. One of
the other questions is how do we prevent

this from getting messed up like it
happens with PSD2? And for the audience

that don't know about PSD2 - PSD2 is a
Directive that forced banks to open up

APIs to other financial service providers,
which is also interoperability between

platforms, in this case for banking
platforms, which comes with all sorts of

privacy questions that weren't completely
thought through when that legislation came

about. Sorry for having this long winded
intoduction, Christoph, but I think it was

needed for people that don't know what
PSD2 means.

Christoph: It's a good question.
Interestingly, we never used PSD2 or the

Telecommunications Act because both have
interoperability options as those negative

examples we always used as examples that
hey, it's already possible. So you don't

have an excuse to say it's impossible to put
it in the law. What is true is that there

is a lot of mess around it. The question
of how to avoid the mess, it's a question

of.. Netzpolitik again. So the question
of whether policy makers are actually

listening to us or listening to industry
lobbyists. So the ones who raised the

question is absolutely right - there's a
huge risk for every topic we talk about.

Whether interoperability, whether it's use
of control about content, targeted ads,

liability, everything that we believe
should be the law, of course, could be

hijacked, could be redesigned in a way
that it will lead to more problems than

fewer problems. So, indeed, for every
policy question we raise, we need to ask

ourselves, is it worth the fight to risk
opening the box of the Pandora? Do we make

it worse? What we said is on that front -
we are happy to make a pressure and what

we need to do in the next year is to
convince them that we are the right person

to talk to. And that's perhaps a challenge
how to make that explicable to policy

makers. So those who ask the questions, I
think those should help us to come to

Brussels to the parliament and tell MEPs
how it's going to work.

Herald: On that note a question to both of
you. Citizen enforcement, I prefer the

term citizen overuses. Would it be helpful
to push for amendments in the parliament

for at least the targeting points you both
mentioned before? And if so, how?

Eliska: So I guess that I will start just
so Christoph can rest a little. So the

question was whether it would be useful to
push for those amendments? Was that's

right?
Herald: For amendments that cover the

targeting of citizens.
Eliska: Absolutely. So there is, of

course, short and long answer as to every
question. And so the short answer would be

yes, but given that the wording of such an
amendment will be precise and nuanced. We

are still working out our positioning on
online targeting and I think we all.. no

one can name those practices that we don't
want to see being deployed by platforms

and where we can actually imagine a proper
ban on such practices. We have recently

published one of our blog posts where we
actually unfold the way of thinking that

Access Now currently, you know, how we
are brainstorming about this whole issue.

And as I said, especially those.. that's
targeting that uses behavioral data of

users, citizens, then maybe let's go for
individuals because they are also obliged

to protect the rights of individuals that
are not their citizens. So that's

definitely one form where we can
definitely see and will be supporting the

ban and possible phase out that will
actually lead to a ban. The same goes for

the cross side tracking of users due to
the fact how users data are being abused

again as being the integral part of the
business models of these platforms and so

on and so forth. So that's one of the
direction we will be definitely taking.

And again, we are inviting all of you to
help us out, to brainstorm together with

us, to assess different options,
directions that we should take into

consideration and not forget about. But I
personally think that this will be one of

the main battles when it comes to DSA,
where we will definitely need to be on the

same page and harmonize and joint the
forces, because DSA gives us a good ground

at the moment, but it doesn't go far
enough. So, yes, definitely the answer is

yes, but given that, we will have a very
nuanced position so we know what we are

asking for and we are taking into
consideration also those other aspects

that could eventually play out badly in
practice. So good intentions are not

enough when it comes to DSA.
Herald: Thank you. I think we are slightly

over time, but I've been told beforehand
it's OK to do so for a few minutes and

there's three questions that are open. One
of one of them, I will answer myself. That

is basically: have the member states
responded? And the answer to that is no.

The member states have not taken any
position. And two others, I think are

quite interesting and important from
a techy perspective. One is: is there

anything you see that might affect current
decentralized platforms like the

Fediverse, Mastodon? And the other is:
will any review of the data protection,

sorry, the Database Protection Directive
affect make the search engines and

interact of these again?
Christoph: Perhaps I jump in and Eliska,

you take over. First, member states have
given an opinion actually on DSA, there

have been one-two official submissions,
plus joint letters, plus discussions in

council on whether the DSA was presented.
I have some nice protocols, which showed

the different attitude of member states
towards it. So for us, it also means we

need to work straight away with the
council to ensure that the package would

be good. What was the question? Ah, yes, I
think the answer to the question depends

on the what lawyers call materials scope
applications whether it would apply to

those platform models at all. But
Eliska you can help me out here. You have

always criticized for the e-Commerce
Directive that it was not quite clear how

it would relate to first non profit
platforms. And many of these alternative

platforms are like that, because there was
this issue of providing a service against

remuneration and what's not quite clear
what it means. Would it apply to Wikipedia

if you get like donations would have
applied to a blogger if you have, like,

pop ups, ads or something like that. So
that's, I think, one huge question. And

the second question is, in as much those
new due diligence obligation would force

alternative platforms, governance models
to redesign the interfaces. And those are

open question for us. We have not analyzed
that in detail, but we see that.. we have

worried that it would not only impact the
large platforms, but many others as well.

What do you think Eliska?
Eliska: Yeah, I can only agree and

especially regarding the non profit
question. Yeah, this is also or was and

always been one of our main asks for
nonprofit organizations. And actually it's

not quite clear how that will play out now
in practice. By the way, how DSA standing,

because at the moment it actually speaks
about online platforms and then it speaks

about very large online platforms, but
what it will be and how it will impact

nonprofit organizations, whether these are
bloggers or organizations like us civil

rights organizations that remain to be
seen. I also know that the European Court

of Human Rights and its jurisprudence try
to establish some principles for the

nonprofit since Delfi AS versus Estonia
and then with the empty e-decision and the

Swedish case that followed afterwards. But
I'm not sure how well it was actually

elaborated later on. But this is something
we will be definitely looking at and

working on further and regarding the
impact on the smaller players and also at

the interface idea, this is still
something we are actually also wondering

about and thinking how this will turn out
in practice. And we are hoping to actually

develop our positioning further on these
issues as well, because we actually

started working, all of us on DSA and on
our recommendations already. I think a

year ago, maybe a year and a half ago when
we were working just with the leagues that

Politico or other media platforms in
Brussels were actually sharing with us.

And we were working with the bits and
pieces and trying to put our thinking

together. Now we have the draft and I
think we need to do another round of very

detailed thinking. What will be ultimately
our position or what will be those

recommendations and amendments in the
European Parliament we will be supporting

and pushing for. So it's a period of
hard work for all of us. Not mentioning

that I always say that, you know, we stand
against a huge lobby power that is going

to be and is constantly being exercised by
these private actors. But I also have to

say that we had a very good cooperation
with the European Commission throughout

the process. And I think that I can say
that on behalf of all of us, that we feel

that the European Commission really listen
this time. So, yeah, more question marks

than answers here from me, I think.
Herald: This is fine with me, not knowing

something is fine. I think we're
definitely run out of time. Thank you both

for being here. Well, enjoy the 2D online
world and the Congress. Thank you. That

wraps the session up for us. Thank you all.
Christoph: Thank you.

Eliska: Thank you very much. Bye.

music

Subtitles created by c3subtitles.de
in the year 2021. Join, and help us!