<html>
<head>
<meta http-equiv="content-type" content="text/html; charset=utf-8">
</head>
<body bgcolor="#FFFFFF" text="#000000">
Please find attached a new draft, 0.96, of the Manila Principles,
for your further comment. For now, to keep things manageable and to
ensure that comments are reviewed widely, we are asking that you
provide your further comments via this list (but if you do want to
be able to comment on a Google Docs version, let us know).
<p>Most of the small changes have simply been accepted because
nobody had objected to them, and they seem objectively reasonable
and consistent with the rest of the document. These are not all
listed here, but there is a record of them in comparing the marked
up text of version 0.9 which remains online. A couple of the more
notable examples only include:</p>
<ul>
<li>
<p>It was pointed out that we should be careful to avoid
allowing intermediaries to restrict content based on their own
terms of service without clearly differentiating between
access providers who are mere conduits, and content hosts. In
the new text of principle I(d) the ability to restrict content
based on terms of service is limited to hosts.</p>
</li>
<li>
<p>It was suggested that we propose that new intermediary
liability rules should require renewal after a five year
period, following an impact assessment—you can see this text
at the very end of the document.</p>
</li>
</ul>
<p>A few of the more extensive changes, although not objectionable
in themselves, would increase the length and complexity of the
principles noticeably, and so for some of these, the suggestion is
made that rather than being included in the principles, they be
included in the part of the background paper that is linked to the
principles. These include:</p>
<ul>
<li>
<p>The introductory paragraph, “The networks that constitute the
Internet are able to interoperate thanks to an architecture
that requires intermediaries to exchange information
automatically, without discriminating between different
messages. Every regulatory burden on intermediaries increases
their costs, making the Internet less accessible to poor
people in industrialized countries and most people in the
developing world.” Also, someone commented that having this
paragraph appear so prominently in the principles seems to be
giving undue deference to the limitation of regulatory burdens
on intermediaries.</p>
</li>
<li>
<p>The comment, “There is a trend towards finding the
intermediary was actually the primary actor (see eg the cases
of Optus TV Now in Australia; Aereo in the US), making any
intermediary protections moot in those cases. Should we have a
principle addressing the trend?” Rather than having a
principle addressing this, we do address it in the background
paper, and can ensure that we do so more thoroughly as we
revise it.</p>
</li>
<li>
<p>The comment, “Shouldn't we come up with proposed standardized
content restriction policies?” This is a good idea! But it
exceeds the scope of this part of the project.</p>
</li>
<li>
<p>A reference was made to the OTI report on transparency
reports at <a
href="http://oti.newamerica.net/blogposts/2014/this_week_transparency_reporting_for_beginners_in_san_francisco_and_austin-104618">http://oti.newamerica.net/blogposts/2014/this_week_transparency_reporting_for_beginners_in_san_francisco_and_austin-104618</a>,
which we will ensure we integrate into the background paper.</p>
</li>
<li>
<p>The suggestion was given to define “least restrictive
technical means”. This too is something that can probably be
best done in the background paper.</p>
</li>
</ul>
<p>Some changes that people had suggested were disputed by others,
and a judgment call has had to be made about whether to include
these or not. Of course, that judgment remains open for discussion
by the broader group. These include:</p>
<ul>
<li>
<p>In “Any rules governing intermediary liability must be laid
down in statute, which must be relevant, accessible,
unambiguous and meet other human rights standards”, there was
a fairly extensive discussion about the inclusion of “other”,
with one person proposing the addition, but two others
opposing it. We are proposing to omit it.</p>
</li>
<li>
<p>Whether search engines are intermediaries. We are proposing
that they are, in line with the UNESCO report and other
documents that we have cited in the background paper.</p>
</li>
<li>
<p>Whether we should talk about counter-notice processes. This
implicitly acknowledges that the legitimacy of a regime that
allows content to be removed on the mere issuance of a notice.
So the current draft does not talk about a “counter-notice”
process in those terms, but does talk about the need for
mechanisms of appeal (against content restriction orders) and
review (against terms of service removals).</p>
</li>
</ul>
<p>Some parts of the text were generally controversial, but a
specific replacement text was not agreed. These include the
following:</p>
<ul>
<li>
<p>That “intermediaries should be allowed to charge for the time
and expense associated with processing legal requests issued
by private parties”, to which we have added, “on a cost
recovery basis”. In my own personal opinion, this is a
paragraph that could be deleted if it continues to be too
controversial, because it is not a core point. What do others
think?</p>
</li>
<li>
<p>What should we say—if anything—about penalties against
intermediaries being “proportionate”. The current text reads,
“Penalties imposed by governments, and damages awarded to
complainants or content providers, against intermediaries for
any act or omission related to third-party content restriction
must be reasonable and proportionate, and in particular
criminal liability must not be imposed.” Again, perhaps, if
this paragraph remains controversial, we could just delete it.</p>
</li>
</ul>
<p>There was division about whether to divide the principles into
two sections or not. Different opinions have been expressed on
this, with Prof K S Park making the case against dividing it. We
therefore no longer have a rough consensus to split the document.
So for now, we have returned the document to a single list of
principles. But to make it easier to see how the principles are
applied in particular liability regimes, we will write a new
non-normative FAQ to accompany the principles. This
is the current proposal, but it can be further discussed.</p>
<p>Finally, there are some more philosophical differences that have
emerged, which we have to make a collective judgment about. These
include some areas on which there is unlikely to be a full
consensus at the end of the day, but we have to aim to achieve the
broadest consensus that we can:</p>
<ul>
<li>
<p>The concept of a “judicial or equivalent order”. Some
expressed reluctance to settle for anything other than a court
order, whereas others point out (per Rishabh Dara) that "we
also need to recognise the need for an institutionalised
redressal mechanism that is cheap and fast for the genuinely
aggrieved.” The text as now presented to require a judicial
order only, as an ideal principle. But this remains open for
discussion, and at a minimum, we will acknowledge in the
background paper that there may be practical difficulties in
reaching this aspirational level and that the use of an
independent administrative process may be the only pragmatic
alternative in some cases.</p>
</li>
<li>
<p>There is a division between those who think that
intermediaries should never be able to manage content on their
networks without oversight from a public authority, and those
who think that they should be able to. For example, Martin
Husovec wrote “I come back to my comments on the background
paper, there you also mention ToS based take-downs; does this
mean that you suggest that if copyright violations are
prohibited, that its OK to go without the court decision, as
long as transparency, etc. is warranted?” As it stands, the
principles do indeed allow an intermediary to act on their own
terms of service, provided that they comply with requirements
of transparency, accountability and procedural fairness.</p>
</li>
<li>
<p>Conversely, some have the view that the principles are
actually too limiting of intermediaries acting without a court
order, and suggest that we should not be so stringent in
requiring that. So for example the case was made by the Web
conference Annemarie Bridy said that notice and takedown
remained appropriate in some cases. In its current format, we
are strongly against notice and takedown, in favour of notice
and notice, or notice and judicial takedown. So this is also a
point of contention. Have we struck the right balance?</p>
</li>
<li>
<p>The suggestion to add “Intermediary liability policies should
be created only after evidence based processes have
demonstrated widespread persistent harms that cannot be
effectively addressed in other ways in the specific
jurisdiction in which the policies will be developed. ”
conflicts with our starting point that there should be a
default policy excluding intermediaries from liability for
third-party content. One comment was “not sure about this”, so
this text has not been integrated, until a better case can be
made for it.</p>
</li>
<li>
<p>On the paragraph “Governments must not impose an obligation
on intermediaries to maintain the ability to de-anonymize
users or identify past user activities, such as by logging
information necessary for such purposes”, it was commented
that disclosure of user identity can be required by the human
rights of victims. Do we want to avoid talking about data
retention in this document? Perhaps this is another point that
is less core to the document, that we could take out if it is
causing too much division.</p>
</li>
</ul>
<p>I hope you find this commentary useful. This summary does not
explain how each and every comment was taken into account, so if
you submitted a comment and cannot easily see how or whether it is
reflected in the new draft, please write to us for clarification.
Note that the background paper has not yet been updated to accord
with the changes to the principles.<br>
</p>
<title></title>
<meta name="generator" content="LibreOffice 4.4.0.3 (MacOSX)">
<style type="text/css">
@page { margin: 2cm }
p { margin-bottom: 0.25cm; line-height: 120% }
a:link { so-language: zxx }
</style>A final note: registration for the Manila Principles meeting
closes on 22 February, which is this Sunday. If you haven't
registered, please do so now! You can do so using this form:<br>
<br>
<a class="moz-txt-link-freetext" href="https://docs.google.com/forms/d/1fP8G1H1xCK0jxkSetI1QtQ7SiCtP508N2xrhwf6RXp0/viewform">https://docs.google.com/forms/d/1fP8G1H1xCK0jxkSetI1QtQ7SiCtP508N2xrhwf6RXp0/viewform</a><br>
<br>
<pre class="moz-signature" cols="72">--
Jeremy Malcolm
Senior Global Policy Analyst
Electronic Frontier Foundation
<a class="moz-txt-link-freetext" href="https://eff.org">https://eff.org</a>
<a class="moz-txt-link-abbreviated" href="mailto:jmalcolm@eff.org">jmalcolm@eff.org</a>
Tel: 415.436.9333 ext 161
:: Defending Your Rights in the Digital World ::
Public key: <a class="moz-txt-link-freetext" href="https://www.eff.org/files/2014/10/09/key_jmalcolm.txt">https://www.eff.org/files/2014/10/09/key_jmalcolm.txt</a>
PGP fingerprint: FF13 C2E9 F9C3 DF54 7C4F EAC1 F675 AAE2 D2AB 2220
OTR fingerprint: 26EE FD85 3740 8228 9460 49A8 536F BCD2 536F A5BD
Learn how to encrypt your email with the Email Self Defense guide:
<a class="moz-txt-link-freetext" href="https://emailselfdefense.fsf.org/en">https://emailselfdefense.fsf.org/en</a></pre>
</body>
</html>