EPDP policy issues - [adding Kathy]

[Raphael Beauregard-Lacroix]

I've always found the natural/legal distinction to be doubtful in this
context, but still I prefer Milton's position, maybe with a slight
modification.

The determination of what is and isnt personal data will be in the hands of
DPAs or judges, at the end of the day. As it has been noted several times,
the definition we're looking at is a relative one: one piece of data may be
personal data in some context, and not in others; and may even be personal
data in the hands of some persons, and not others... the key being whether
it allows the identification of a natural person.

Hence for the purpose of the scope of the law and its operation, it does
not matter whether one says "there is no personal data in these form
fields" or "I am not a natural person but a legal person." Not that people
are idiots and cannot understand the law, but their determination is not
binding on anyone and does not have legal effect. For example, if I claim
to be a legal person (hence out of scope of the law) and proceed to give
someone else's full address+name+phone number, the fact that this data may
genuinely relate to a legal person, and that I made that claim by ticking a
box, does not change the fact that the disclosure, in this case, amounts to
processing of personal data, which puts me "back" into the scope of the law.

It does matter, however, whether someone says "I consent to this processing
of data *about me*" (here would be the modification) That makes the
processing legal, provided that the general rules on consent and the
provision of information about the processing are followed.

Consent as a basis is fraught with issues, but even for those like me who
are generally skeptical of it, I think that the processing and the data
concerned are narrow/limited enough to make consent an otherwise
"acceptable" legal basis for processing.

Now if in consenting, people end up disclosing personal data pertaining to
someone else, willingly or unwillingly, with or without "bad" intent, well
yes we do have a problem... and ticking an additional box will not absolve
them of liability under their local law (or any other), as the case may be.
If that is such a potential issue, however, then the only solution that I
see is not to have any disclosures possible outside of mandatory ones. But
that would be out of scope, as I understand it, and as pointed out by
Manju, and would raise the potentially valid point of what if someone wants
to disclose.

Additionally, I do not think that forcing mandatory rules or refraining
from saying anything would be a good idea at this point.

Have a nice day,



On Wed, Apr 28, 2021 at 10:53 AM Stephanie E Perrin <
stephanie.perrin at mail.utoronto.ca> wrote:

> Thanks again Manju, for your untiring efforts to get us to solve this
> disagreement about next steps.  I am going to try to summarize my thinking
> on this, without diving into the complexity of the law and the difficulties
> inherent in implementing it.
>
> 1.  The push to distinguish between legal and natural is not new.  We
> fought it in the PPSAI, and won.  In phase 1 EPDP, we managed to get
> recommendation #6 through, but with the concession that ICANN would do a
> study on legal natural, and that we would reexamine.  My position is
> basically HOLD THAT LINE!  We do not need to change our position, we can
> examine the matter, take a look at what guidance would do, and maintain the
> status quo.
>
> 2.  We have sought advice from Bird and Bird on mitigating the risk to
> contracted parties in terms of how they provide information to their
> registrants, in order to be in full compliance with data protection law
> (viewed through the GDPR lens).  For the contracted parties, nothing in
> there eliminates the risk that the data controller has, but certainly
> guidance mitigates it.  However, remember that civil society can take a
> case under GDPR, and if I were advising civil society as to how to take a
> case, I would point out the history of WHOIS, the dogged determination of
> ICANN and the IP and business community to get back to the good old WHOIS,
> the lack of evidence that forcing this determination on smaller actors will
> indeed result in greater security and stability  of the DNS, and the
> unequal power relationship between ICANN the regulator and the contracted
> parties who must be accredited to do business.  Third party data accessors
> are driving this process, and registrants are basicly being represented by
> ourselves, and the Registrars who have them as customers.  I think we have
> a responsibility not to cave in.
>
> 3.  A word on small business, sole proprietors, and home based
> entrepreneurs or gig workers.  We are the Non-commercial stakeholders, so
> we do not claim to represent them.  Frankly, I don't know who does here at
> ICANN, in my 8 years of volunteering at ICANN I have never heard the ALAC
> folks advance a cohesive argument on behalf of these folks (or even an
> argument, but I hesitate to say that because someone will trawl through the
> archives and come up with someone defending the little guy against giant
> corporations).  Certainly they are not speaking for them at the EPDP, they
> are speaking for government, law enforcement, and cybersecurity
> operatives.  As folks who care about human rights and fairness to
> developing economies, I think we should care about how much this
> differentiation between legal persons and natural persons does not work in
> countries other than the EU states and the US.  Countries around the world
> have different ways to describe small business, different ways to regulate
> it, and this differentiation may not match tax schemes, municipal
> registration patterns, etc.  There is of course an additional barrier in
> the matter of languages.
>
> 4.  Remember that what we are arguing about is not providing access to
> data about suspected legal persons through the SSAD.  We are arguing about
> Automatic disclosure based on the choice the individual makes, legal or
> natural person.  As I have said before, the skies are not going to fall if
> the legitimate requestors have to request the data and get it in one or two
> business days, as opposed to having it pre-emptively disclosed.   Remember
> that the registrar or his reseller has a wealth of other day about their
> customer that is "below the surface", most importantly credit card info,
> billing address for the credit card, emails, IP address etc.  They can look
> at that data if necessary, to figure out whether they are dealing with a
> company or a person.  However, the SSAD does not have access to that data.
> If we encourage this differentiation by putting guidance into the policy,
> then we are heading for the slippery slope of automated disclosure.
> Remember that there is already a section in there that once the data has
> been verified as pertaining to a legal person and not containing personal
> data, it MUST be disclosed.  What happens when that legal entity moves to a
> jurisdiction where employees have privacy rights, either under data
> protection law or other laws such as labour laws or contracts?  There is a
> good chance that some of their data becomes personal.
>
> 5.  Registrars already have excellent advice for their members available
> through their own websites.  Remember that at ICANN we are only dealing
> with the big ones, and the responsible ones.  We have to consider whether
> any activity we sanction in this policy induces the lazy ones to cut
> corners.  They are used to disclosing everything in the WHOIS, if given an
> opportunity to revert to that by creating this distinction, they will most
> likely do what is easiest.  At the moment, under the temp spec and article
> 6, the easiest thing to do is to consider these grey area folks as natural
> persons unless proven otherwise, and protect the data.  Why on earth would
> we not choose this option?  There is no law telling us to do otherwise, and
> there is certainly a great deal of law out there that makes differentiation
> a legal risk that carries liability and cost.  We want domain names to
> remain affordable, and we want our people protected.
>
> 6.  One more thing:  remember that not all domain names are used for
> websites engaged in commerce.  Some are being held by individuals for
> future use.  Disclosure provides market information to big players who may
> want to prevent the name from being used, or to purchase them, but there is
> no harm emanating from them in dormancy, and small players need not have
> their competitive positions compromised in this way.  Most of the online
> crime that we hear adduced to justify disclosure of data is coming from
> websites.  Websites can be regulated to protect consumers, but that is not
> within ICANN's bailiwick.  Folks come to ICANN to demand this activity
> because other multilateral instruments have failed, but given the lack of
> oversight over ICANN, the responsibility to ensure fairness and human
> rights are respected falls on the shoulders of the stakeholders engaged
> here.
>
> 7.  Finally:  Several folks are worried that governments will regulate if
> we do not concede here.  I say let them, the Courts and the Constitution
> and the much more rigorous defences against incursions into human rights
> available within countries will prevail.  It certainly will in the EU,
> where the draft NIS is progressing.....I keep pointing out the Opinion of
> the European Data Protection Supervisor on this matter, so one more time
> here it is
> https://www.google.com/search?client=firefox-b-d&q=edps+on+the+nis+directive.
> As for the increasing number of autocratic countries interested in
> curtailing the free speech and internet rights of their citizens, they are
> regulating anyway and nothing that ICANN does is going to persuade them
> otherwise.  A good example of us falling on the side of customer protection
> might at least give those individuals in these countries something to point
> to, as opposed to caving in to threats.
>
> cheers Stephanie Perrin
> On 2021-04-27 11:56 p.m., 陳曼茹 Manju Chen wrote:
>
> *EXTERNAL EMAIL:*
> Hi all,
>
> This has been a very helpful discussion. Our ultimate goal is to finalize
> an NCSG position, and I'd like to provide some context in the interest of
> reaching that goal.
>
> Some have pointed out that the question of 'do you agree to publish your
> data' is a much better one than 'are you a legal or natural person'. I
> totally agree. But we already have that in the policy. In the EPDP phase 1
> final report, recommendation #6 reads:
>
>> The EPDP Team recommends that, as soon as commercially reasonable,
>> Registrar must provide the opportunity for the Registered Name Holder to
>> provide its Consent to publish redacted contact information, as well as
>> the email address, in the RDS for the sponsoring registrar.
>
>
> In the EPDP phase2A, the WG is tasked to address the following questions:
>
>    1. Whether any updates are required to the EPDP Phase 1 recommendation
>    on this topic (“Registrars and Registry Operators are permitted to
>    differentiate between registrations of legal and natural persons, but are
>    not obligated to do so“);
>    2. What guidance, if any, can be provided to Registrars and/or
>    Registries who differentiate between registrations of legal and natural
>    persons.
>
> So it's not that we 'want' the distinction of natural vs. legal. It's the
> task we're given, and per the WG chair's last email to the EPDP team, it's
> not something we can avoid. We have to stick to the 'natural vs. legal
> persons' rhetoric to answer the questions.
>
> Of course, we can simply assert that 'there's no update needed' and let
> the contracted parties do whatever they want when they feel like making the
> distinction. I was personally very tempted by this option every now and
> then. However, as Milton pointed out in another email, we are already
> developing guidance in the EPDP, so that option seems far-fetched.
>
> I hope this explains some of the confusion and helps us to consider
> Milton's 4 options in the other email.
>
> Thanks!
>
>
> Best,
> Manju
>
> On Wed, Apr 28, 2021 at 4:40 AM Mark Leiser <markleiser at gmail.com> wrote:
>
>> Hi Milton,
>>
>> First let me say I completely agree with you on your statements about
>> European data privacy. I feel like I've dedicated most of my professional
>> life arguing about the dangers of privacy and data protection maximalism -
>> yet almost always feel like my arguments on deaf ears. While people are
>> arguing that 'everything is personal data', I've been arguing that this
>> makes the regime unmanageable.
>>
>> So let me try to explain 'relating to' with reference to 'Milton
>> Mueller's Porkbelly Diner'. Because of Recital 14, this would amount to a
>> legal person. It's pretty clear that the intention of the GDPR's drafters
>> was to exclude legal persons. However, let's say you have registered 'MM
>> Porkbelly Diner' in the register of companies. I think you would agree that
>> this would amount to information about a legal person. But say
>> someone searched the company register and discovered Milton Mueller was the
>> principal shareholder of MM Porkbelly Diner. This is an identifier which
>> would *tie nformation about Milton Mueller to* 'MM PorkBelly Diner'.
>> Therefore, this is 'any information' 'relating to' an identified or
>> identifiable living person. It would be reasonable to infer that MM
>> registered in the registrar database under MM PorkBelly Diner is the same
>> as the Milton Mueller that is in the register of companies. The fact that
>> someone can combine the knowledge from the company register with the
>> knowledge from the registrar database could make 'MM Porkbelly Diner'
>> personal data under Article 4(1) of the GDPR. Someone who did not disclose
>> their identity at all could still be identifiable; hence, the perceived
>> need for protection in the EU data protection regime.
>>
>>  "I am saying that the user, the registrant, gets to decide what is
>> personal data or not, because THERE IS NO OBJECTIVE, CLEAR LEGAL
>> DEFINITION" is, on the surface, problematic. It doesn't matter whether
>> the user says NO or YES or the registrar says no or yes, or whether it is
>> objective or clear, the test is whether any information can be combined
>> with other information to reveal an identifiable living person. I would
>> argue, absent a wholesale change in the reasoning used by the CJEU, this
>> would remain the case for the foreseeable future. This is not intended as a
>> Mark Leiser argument or an attempt to discredit what you are saying, but an
>> honest account of how I think the Courts and the EU data protection Board
>> would react to what you are proposing.
>>
>> You asked about IP addresses
>> <https://iapp.org/news/a/are-ip-addresses-generated-when-users-visit-websites-personal-information/#:~:text=Under%20the%20EU%20General%20Data,header%20information%20that%20website%20hosts>,
>> 'port numbers', 'browser config', etc could be used to identify you
>> personally. Yes, absolutely. This is personal data in the EU - if it can
>> relate to a living person. What your writing here reveals, is that you are
>> a little confused about how the GDPR works - the GDPR does not rely on
>> 'consent' or 'explicit permission' as the only basis for processing
>> personal data. Remember the GDPR has six grounds of processing
>> <https://gdpr-info.eu/art-6-gdpr/>. It is a prohibitive regulation. You
>> cannot process personal data in the EU unless you satisfy one of those six
>> grounds. Most companies will NOT be processing on the basis of 'consent'
>> but on 'legitimate interests' (Article 6(1)(f)) or performance of a
>> contract (Article 6(1)(b)). Don't worry, this is an extremely common
>> mistake among American attorneys! As most of these items are "technical', I
>> would also imagine that there is a 'legal requirement' (another ground) or
>> a lawful basis. All ISPs will be processing personal data through 'IP
>> addresses', 'port browsers', and 'browser config' because of the legal
>> basis found in Article 15 of the e-Privacy Directive
>> <https://eur-lex.europa.eu/legal-content/EN/TXT/HTML/?uri=CELEX:32002L0058&from=EN> which
>> provides Member States with a specific exemption for the purposes of
>> national security. If not covered by this, I would imagine they would rely
>> on 'legitimate interests' as their ground instead.
>>
>> Finally, even if a registrar has a legitimate interest in processing
>> someone's personal data, this does not address the *privacy *requirements
>> that i indicated in my previous email. Unless there is a specific provision
>> put into law, I do not know how the registrar can remain compliant with the
>> GDPR and the EU's privacy requirements. In fact, I don't know how the EU
>> can even comment on this, because, of course, the EU Charter is a legal
>> framework completely distinct from the European Convention of Human Rights.
>>
>> One more thing to consider - if you choose to disclose your name,
>> address, and designate as the contact person of a legal person, what
>> happens when that person exercises their 'right to be forgotten' right
>> under Article 17 GDPR. Maybe Milton Mueller the person chooses to leave the
>> legal entity MM Porkbelly Diner, adopts a vegan lifestyle, and wants no
>> affiliation with the business? If it's published, he can demand a
>> correction of the database thereof. So what then?
>>
>> Regards,
>>
>> Mark
>>
>> *Dr Mark Leiser | Law and Digital Technologies | FRSA FHEA |*
>>
>>
>> On Tue, 27 Apr 2021 at 20:35, Mueller, Milton L <milton at gatech.edu>
>> wrote:
>>
>>> Mark,
>>>
>>> Thanks for your intervention. Here is the complete definition of
>>> personal data in GDPR:
>>>
>>> ‘personal data’ means any information relating to an identified or
>>> identifiable natural person (‘data subject’); an identifiable natural
>>> person is one who can be identified, directly or indirectly, in particular
>>> by reference to an identifier such as a name, an identification number,
>>> location data, an online identifier or to one or more factors specific to
>>> the physical, physiological, genetic, mental, economic, cultural or social
>>> identity of that natural person;
>>>
>>> Alas, this raises more questions than it answers. It is not entirely
>>> clear what “relating to” means in this construct. E.g., if the name of my
>>> company is Milton Mueller’s Porkbelly Diner, is the name relating to me as
>>> a person, or to my business? Strictly speaking it’s a business name. But it
>>> could be used to identify me. An overly broad interpretation of this
>>> definition would classify ANY data about ANYTHING as “personal data”
>>> because at some point it could be “related to” an “identifiable natural
>>> person.” So suddenly the business name becomes personal data. Or a bunch of
>>> obscure technical indicators inherent in your use of the internet, such as
>>> port numbers, browser config, etc., could be “related” to your ISP account
>>> number, and then used to identify you, personally. But does that mean that
>>> every website and hosting service in the world that uses that technical
>>> data in the course of their operations cannot process that info without
>>> your explicit permission, because it’s “personal data?” I hope not, because
>>> the internet would cease to function if so.
>>>
>>>
>>>
>>> >Therefore, info at myorg.org is personal data if someone behind it is
>>> identifiable.
>>>
>>>
>>>
>>> My point is that whether the user of info at myorg.org is identifiable
>>> does not depend on that chunk of data, but on a bunch of activities that
>>> relate that data to other things. And in a digital world with powerful
>>> processing capabilities, no one can fully control those correlations and
>>> searches. Any attempt to do so simply cripples the entire information
>>> economy. So the idea of looking at a registration record and saying “is
>>> there personal data in here or not” is a completely invalid test.
>>>
>>>
>>>
>>> There is an eerie similarity between these exaggerated applications of
>>> privacy law and the copyright maximalists of the 1990s. The IP interests
>>> thought you needed permission to transmit a copyrighted work over the
>>> internet, a claim that would have crippled ISPs who had no idea what
>>> packets were parts of copyrighted material or not. Privacy maximalists have
>>> reached the same point of absurdity, but they don’t seem to realize it.
>>> They are no longer protecting a tangible privacy interest of internet
>>> users, they are trying to give individuals rigid control over information
>>> exchanges and imposing largely meaningless consent requirements that do no
>>> one any good.
>>>
>>>
>>>
>>> >The confusion comes from Recital 14 of the GDPR which states that it
>>> only applies to natural persons and does not cover the processing of
>>> personal data concerning legal persons, in particular undertakings
>>> established as legal persons or legal entities. This also includes the name
>>> of the legal person, the form, and the contact details of the legal person. But
>>> the second you start adding identifiers to these details, it stops becoming
>>> the data of a 'legal person' and starts becoming personal data.
>>>
>>>
>>>
>>> Yes, indeed, the GDPR is confused on this score. Unfortunately, you
>>> don’t escape that confusion by saying “adding identifiers” is the problem.
>>> Your argument fails because names of legal persons and contact details ARE
>>> “identifiers,” they are just identifiers of legal persons. As I pointed out
>>> they can easily overlap with, or be used to identify, natural persons.
>>>
>>>
>>>
>>> So my basic point is that we do not solve this problem by reference to
>>> GDPR definitions. In fact if GDPR is taken literally no one can ever
>>> publish and share any kind of information without absurd overhead and legal
>>> bureaucracy, because ALL of it can be used to identify you in some way.
>>> These kinds of interpretations actually discredit privacy laws and
>>> protections, by taking them to counterproductive lengths. We have to make
>>> common sense-based, practical distinctions between what data needs
>>> protection, what data can be easily shared at the registrants’ choice.
>>>
>>>
>>>
>>> Secondly, I find this statement, "I think it is simplest to just say to
>>> the registrant, if you’re a company and don’t mind (or even want) your data
>>> to be published, check this box. If you’re not, or you do mind, don’t check
>>> it."  to be BOTH perplexing and inaccurate.  Whether someone checks the
>>> box, or does not check the box is completely irrelevant for determining
>>> whether it is personal data. Whether someone checks the box is a question
>>> of CONSENT to processing AND publication, not whether this amounts to
>>> personal data.
>>>
>>>
>>>
>>> We actually seem to agree here, more than you think. I am saying that
>>> the user, the registrant, gets to decide what is personal data or not,
>>> because THERE IS NO OBJECTIVE, CLEAR LEGAL DEFINITION. So the user can
>>> decide whether they want to be classified as a legal person and “consent”
>>> to publishing their data or not. As an example, that if I decide that
>>> publishing the name of “Milton Mueller’s Porkbelly Diner” is not a
>>> violation of my privacy, it isn’t. It doesn’t matter what the European
>>> Union says, it’s my choice.
>>>
>>>
>>>
>>> And let’s not overlook the embarrassing fact that the European Union is
>>> now one of the key players pushing hard for publication of legal person
>>> data. But I’ll leave that one to later.
>>>
>>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ncsg.is/pipermail/ncsg-discuss/attachments/20210428/436aeff2/attachment.htm>