[GNSO-Accuracy-ST] Potential Additional Questions to ICANN Org

Wed Mar 2 14:10:19 UTC 2022

Hello Becky,

In my capacity as the Chair, I view my responsibilities as trying to balance
the respective perspectives of all participants and making sure that we
document all of these perspectives in our work product.  As noted before, I
would like to avoid some of the holes identified in the SSAD ODA, by making
sure that we look at all sides of issues (both popular and unpopular).

I think most will agree with me, that there have been some deeply held and
divergent viewpoints that were crystal clear at the start of this Group's
work, see
https://mm.icann.org/pipermail/gnso-accuracy-st/attachments/20211005/46ce22c
7/AccuracyScopingTeam-InitialInput-5October2021-0001.pdf  I think it is fair
to characterize the Registrar's position as follows: the 2013 RAA provides a
black and white guidance as to the term accuracy; recent ICANN Compliance
reports indicate that there are a de minimis number of accuracy complaints;
therefore there is no problem let's move on. I also think it is fair to
characterize the BC, IPC, ALAC, SSAC and GAC respective positions as the
2013 RAA definition is over narrow; the recent ICANN Compliance reports are
under-reporting inaccuracy because legitimate third party access has largely
gone dark, therefore we need to reinitiate some type of accuracy survey to
see the scope of the problem (if any) and propose a new definition (if
necessary).

So as Chair, I started out with the proposition that the 2013 RAA was a
rebuttal status quo baseline for defining accuracy. Initially, there were
several groups that rebutted that definition.  In fact, the push back was so
hard we did not even use the term "definition" for a couple of weeks.
However, after ICANN Compliance provided their feedback it was clear that
the scope of accuracy was not simply a 2013 RAA black and white syntactical
and operational exercise.  My job as Chair was then to help define the scope
of that "grey" for purposes of our working definition as well as potential
studies that might be undertaken to help document any accuracy problem. 

In response to calls from a number of members that wanted to restart the
ARS, which was unilaterally suspended by ICANN Org after the GDPR went into
effect, my job as Chair was to help find a potential middle ground if
possible. While being respectful of contracting parties data privacy
concerns, including the lack of no DPA with ICANN, I wanted to find the
optimal path forward for a potential new ARS that would be in compliant with
Article 6.1.(f) of the GDPR: 

processing is necessary for the purposes of the legitimate interests pursued
by the controller or by a third party, except where such interests are
overridden by the interests or fundamental rights and freedoms of the data
subject which require protection of personal data, in particular where the
data subject is a child.

So instead of processing data from a universe of over 200 million gTLD
domain registrations, I thought ICANN Org would be able to maximize its
legitimate interest (while also taking into account the interests of the
registrant) by focusing on a much narrower class of domain names associated
with documented DNS abuse. Now there is almost universal recognition that
there is a DNS Abuse problem, see DNS Abuse Institute, topDNS, Global Cyber
Alliance, EU DNS Abuser Report, and CoCCA DSI. In fact the Registries and
Registrars are both supportive of the upcoming DNS Abuse session that will
undertake an analysis between maliciously registered domain names and
compromised domain names.  

What I find odd is the almost universal support of Registries and Registrars
to look at a DNS Abuse subset of domains to make a "malicious / comprised"
determination, yet the mere suggestion that this same subset of data could
be analyzed through the lens of "accuracy" is somehow verboten. I think you
would agree with me that to an objective outside observer, the Registrar
pushback appears less likely tied to statistical purity and potentially more
likely because of what those results may reveal.  To those Registries and
Registrars that may be participating in the malicious v compromised domain
name study, I assume you are contacting Registrants. Would it be that much
of an additional burden to ask the Registrants if the information contained
in the Whois/RDDS is accurate and share those results with the group?  

In respond to your comment about "understand[ing] the interest in asking
ICANN about its legal advice, but I am skeptical that will prove
particularly enlightening." Perhaps if ICANN Legal just provided a legal
basis to the community instead of making self-serving proclamations about
its role that would help out from a trust building exercise.  These are the
concerns that I heard on the call last week when Thomas raised the issue
about the lack of a DPA and Stephanie's repeated request for ICANN to file a
DPIA over the past several years.  Additionally, there have been other
members that have raised similar concerns, so as Chair I will continue to
push for these voices within the Working Group to be heard and to obtain the
data for this group to make a factual determination.

Best regards,

Michael

From: Becky Burr <BBurr at hwglaw.com> 
Sent: Tuesday, March 1, 2022 1:29 PM
To: 'STROUNGI Melina' <Melina.STROUNGI at ec.europa.eu>; 'Becky Burr'
<becky.burr at board.icann.org>; michael at palage.com
Cc: gnso-accuracy-st at icann.org
Subject: Re: [GNSO-Accuracy-ST] Potential Additional Questions to ICANN Org

Re "clearing the bar" for legitimate interests not outweighed by the data
subjects' privacy rights, what I think is probably not particularly
relevant.  In the end, it will fall to the party granting access to the data
(i.e., the Contracted Party) to make that call.  But in general I do agree
that processing based on a specific cause for concern will usually be easier
to justify under GDPR.

That said, I do agree with the CPs who expressed concern about skewed
results.  If we want to actually understand the volume and nature of
inaccuracy across the entire data set and we wanted an answer that cannot be
dismissed by one group of stakeholders or another, then a proactive audit
that looks at registrant data across domains and across sponsoring
registrars is likely necessary.  I'm not a statistics guru, but I suspect
you could design a study that looked at a subset of the data, but looking
only at the subset of data that has already been identified as inaccurate
seems very problematic to me.

I understand the interest in asking ICANN about its legal advice, but I am
skeptical that will prove particularly enlightening.  

While I'd prefer to use the actual language from the 6.1(f) exception
(legitimate interests not overridden by the interests/rights/freedoms of the
data subject) using "under the GDPR" rather than "proportionate" also works.

J. Beckwith Burr

HARRIS, WILTSHIRE & GRANNIS LLP

1919 M Street NW/8th Floor

Washington DC 20036

202.730.1316 (P) 202.352.6367 (M)

  _____  

From: GNSO-Accuracy-ST < <mailto:gnso-accuracy-st-bounces at icann.org>
gnso-accuracy-st-bounces at icann.org> on behalf of Michael Palage <
<mailto:michael at palage.com> michael at palage.com>
Sent: Tuesday, March 1, 2022 12:14:56 PM
To: 'STROUNGI Melina'; 'Becky Burr'
Cc:  <mailto:gnso-accuracy-st at icann.org> gnso-accuracy-st at icann.org
Subject: Re: [GNSO-Accuracy-ST] Potential Additional Questions to ICANN Org 

Hello Becky,

I think we are in agreement that the processing/disclosing of non-public PII
involves a two part test: legitimate interest and proportionate interest.
As you and others may recall this is why I have proposed that any limited
restart of the ARS program involve a sampling of the data from the monthly
DAAR reporting. 

>From a legitimate interest standpoint, the domains reported in DAAR (e.g.
malware, phishing, SPAM) are clearly involved in illegal activity in most
jurisdictions. I do not see any situation in which ICANN would not easily
clear this bar. Would you agree?

With regard to the proportionality (balancing test), as will be discussed in
the upcoming ICANN73 meeting, there are two types of potential registrations
involving abusive domains, maliciously registered domain names and
compromised domain names. From a "balancing test" ICANN easily clears any
proportionality bar when looking at maliciously registered domain names.
Would you agree? 

With regard to compromised domains, while this balancing test is a little
more substantive than with malicious domain names, I believe this is a bar
that ICANN Org and the Contracting Parties should easily be able to clear in
almost every scenario. Unlike the old Whois/RDDS that made registrant data
publicly available for scraping, this proposed audit would be limited to a
restrictive number of parties, ICANN, third-party vendor, and contracting
party. Would you agree?

I am also in agreement with your comments distinguishing between targeted
processing/disclosure versus bulk processing/disclosure.  This is why I made
the specific proposal to restart ADR on a limited scale targeting just
Abusive Domain Names reported via DAAR. This targeted focus should address
your bulk processing claims. Would you agree? 

What I found disappointing when I brought this to the consideration of the
entire Working Group is that several Contracting Parties opposed this
potential reasonable path forward because they thought that this would
potentially skew the accuracy results. These Contracting Parties instead
were adamant that any survey would need to involve the entire data set. My
concern with this position is that any demands to include the entire set is
potentially a non-start for processing in a legal compliant manner per the
GDPR. 

In order to further explore your claim that the DPA is a red hearing,
perhaps the Registry and Registrars representatives could go back to their
respective stakeholders groups and ask for scenarios in which they would be
willing to transfer data to ICANN or a designated vendor to check the
accuracy of data. Would you agree with me that this data point would be
extremely helpful in resolving potential ambiguity between the parties and
their respective roles?

In closing, I want to thank both Becky and Melina for your respective
feedback and I look forward to additional constructive feedback going
forward.

Best regards,

Michael

From: STROUNGI Melina < <mailto:Melina.STROUNGI at ec.europa.eu>
Melina.STROUNGI at ec.europa.eu> 
Sent: Tuesday, March 1, 2022 11:13 AM
To: Becky Burr < <mailto:becky.burr at board.icann.org>
becky.burr at board.icann.org>;  <mailto:michael at palage.com> michael at palage.com
Cc:  <mailto:gnso-accuracy-st at icann.org> gnso-accuracy-st at icann.org
Subject: RE: [GNSO-Accuracy-ST] Potential Additional Questions to ICANN Org

Dear Becky, all,

many thanks for your additional suggestions.

my two cents on the below:

I support the distinction proposed by Becky, but would recommend replacing
'proportionate' with 'under the GDPR' so we are fully covered.

The original question raised was "Is ICANN able to access registration data
under the GDPR on the basis that it has a legitimate interest in checking
the accuracy of the data? "

The reference alone to the GDPR means that indeed the balancing test has
been taken into account. As you rightly point out the GDPR requires a
balancing test when 'legitimate interests' is used as a legal basis. So in
my view as long as there is a reference to the GDPR there is no need to
explicitly add the proportionate part. It is already implied. Plus, a more
general reference is more encompassing in the sense that it takes into
account the totality of the balancing test (i.e., data subjects' interests
etc.)

Having said that I am all supportive of asking all of these questions (in
general I am in favor of asking as many questions as we can think of as this
is at the heart of our scoping tasks), but I would maintain -on top of what
you suggest -the specific question on whether ICANN ever received or plans
to receive legal advice on this particular topic.

If I recall correctly this had been discussed in our accuracy scoping
meeting  of 17 February and was proposed as a question to be addressed to
Brian so he can forward it to ICANN compliance.

In order to be able to progress with our discussions, it is important to
know where exactly ICANN would base their assessment on these questions
(i.e., whether they have received specific in-house or external legal
advice, including but not limited to any correspondence with the EDPB). 

I have now tried to integrate Becky's suggestion to the original questions.
Hope this helps.

Question 1

1.	Does ICANN have a legitimate interest under the GDPR in accessing
domain name registration data in response to complaints that the data is
inaccurate? Has ICANN ever received or plans to receive legal advice on this
particular topic?  
2.	Does ICANN have a legitimate interest under the GDPR in proactively
acquiring bulk access to domain name registration data to undertake an
accuracy audit, even with respect to data for which it has no basis to
question its accuracy? Has ICANN ever received or plans to receive legal
advice on this particular topic?  

Question 2

For either scenario a or b under question 1: Does ICANN believe that a Data
Protection Agreement between itself and the Contracted Parties is a
necessary legal requirement for requesting and receiving this data, and if
so for what legal reason? What happens if the registrar receiving the access
request disagrees with ICANN's application of the balancing test, i.e., does
ICANN have the contractual authority to enforce its access request? 

Best,

Melina

From: GNSO-Accuracy-ST < <mailto:gnso-accuracy-st-bounces at icann.org>
gnso-accuracy-st-bounces at icann.org> On Behalf Of Becky Burr
Sent: Tuesday, March 1, 2022 4:27 PM
To:  <mailto:michael at palage.com> michael at palage.com
Cc:  <mailto:gnso-accuracy-st at icann.org> gnso-accuracy-st at icann.org
Subject: Re: [GNSO-Accuracy-ST] Potential Additional Questions to ICANN Org

Michael -

Respectfully, and without taking a position on whether these questions are
relevant or timely, I think the questions need to be more nuanced to produce
useful answers.

3.	Does ICANN have a legitimate and proportionate interest in accessing
individual registration records in response to credible complaints that the
data is inaccurate? If so, is a DPA required to access data in such
situations?  What happens if the registrar receiving the access request
disagrees with ICANN's application of the balancing test, i.e., does ICANN
have the contractual authority to enforce its access request? 
4.	Does ICANN have a legitimate and proportionate interest in
proactively acquiring bulk access to registrant data to undertake an
accuracy audit, even with respect to data for which it has no basis to
question its accuracy?  If so, is a DPA necessary to do so?  What happens if
the registrar receiving the access request disagrees with ICANN's
application of the balancing test, i.e., does ICANN have the contractual
authority to enforce its access request?

It is important to keep in mind that a legitimate interest is necessary but
not sufficient under GDPR.  The processing necessary to satisfy a legitimate
interest must be proportionate, i.e., not outweighed by the privacy rights
of the individual data subject(s).  As a result, the two situations (access
to a single record based on reasonable grounds to believe the data is
inaccurate v. proactive access without individualized suspicion) are quite
different from a data protection perspective, with the first being far less
complicated to defend.  In addition, a CP's contractual obligations, e.g.,
under the RAA, may be different in those situations.

FWIW, I think the DPA issue is a bit of a red herring here.   Presumably,
ICANN's requests for one-off data can be handled in the same way that anyone
else's access request is handled, e.g., if the data is to be transferred
outside of the EU by imposing controller to controller Standard Contractual
Clauses as the terms and conditions of such access on a case-by-case basis.
If the EDPB were to confirm that ICANN's bulk access to data for proactive
checking was legitimate and proportionate, it's clear to me that a narrowly
focused DPA between ICANN and CPs applicable to data access for the specific
purpose of checking accuracy (e.g., prohibiting onward transfer, etc.) could
be crafted.  The real question is whether (i) the temp spec /epdp phase 1
policy obligating CPs to provide reasonable access for legitimate and
proportionate purposes encompasses bulk access or (ii) some other provision
of the agreements produces an obligation to provide bulk access.  

Apologies for being pedantic here.  None of us can say with any certainty
what GDPR does or does not permit as that determination is ultimately made
by individual data protection authorities and/or the EDPB.  We are asking
ICANN for its views on what GDPR would permit in specific circumstances, so
the relevant circumstances should be articulated precisely to produce useful
answers.  

b

On Thu, Feb 24, 2022 at 5:12 PM Michael Palage <michael at palage.com
<mailto:michael at palage.com> > wrote:

Hello Everyone,

Over the past couple of weeks there has been a recurring theme in our calls
and in some of the side discussions that I have had with some members
regarding about how the potential lack of a Data Processing Agreement
between ICANN Org and the Contracting Parties might negatively impact our
future work and/or recommendations.

Therefore I would like to propose to the group for their consideration the
following additional questions that we may want to propose to ICANN Org as
we continue our work:

.            "Is ICANN able to access registration data under the GDPR on
the basis that it has a legitimate interest in checking the accuracy of the
data?  Has ICANN ever received or plans to receive legal advice on this
particular topic?  

.            Does ICANN believe that the Data Protection Agreement between
itself and the Contracted Parties is a necessary legal requirement for
requesting and receiving this data, and if so for what legal reason?"  

As always I welcome any thoughts and or considerations?

Best regards,

Michael

_______________________________________________
GNSO-Accuracy-ST mailing list
GNSO-Accuracy-ST at icann.org <mailto:GNSO-Accuracy-ST at icann.org> 
https://mm.icann.org/mailman/listinfo/gnso-accuracy-st
<https://urldefense.com/v3/__https:/mm.icann.org/mailman/listinfo/gnso-accur
acy-st__;!!DOxrgLBm!UNIutLhtKN79DDAyXMFbwLBM5YmXBcSEX_Z4GrdQYmzW9hzoy8-8tEvp
-nQ58gfd3kyLTw-r$> 

_______________________________________________
By submitting your personal data, you consent to the processing of your
personal data for purposes of subscribing to this mailing list accordance
with the ICANN Privacy Policy (https://www.icann.org/privacy/policy
<https://urldefense.com/v3/__https:/www.icann.org/privacy/policy__;!!DOxrgLB
m!UNIutLhtKN79DDAyXMFbwLBM5YmXBcSEX_Z4GrdQYmzW9hzoy8-8tEvp-nQ58gfd3uHvF4Qg$>
) and the website Terms of Service (https://www.icann.org/privacy/tos
<https://urldefense.com/v3/__https:/www.icann.org/privacy/tos__;!!DOxrgLBm!U
NIutLhtKN79DDAyXMFbwLBM5YmXBcSEX_Z4GrdQYmzW9hzoy8-8tEvp-nQ58gfd3p4iJdXc$> ).
You can visit the Mailman link above to change your membership status or
configuration, including unsubscribing, setting digest-style delivery or
disabling delivery altogether (e.g., for a vacation), and so on.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mm.icann.org/pipermail/gnso-accuracy-st/attachments/20220302/43f2a4bb/attachment-0001.html>