Direct data and derived data: chance for a data GPL?

Direct data and derived data: chance for a data GPL?

[Andy Oram]

The world is abuzz with debates over user data in the hands of social media
companies and tech corporations. One of the distinctions that observers
note is between data that we provide directly--by posting--and data that is
derived from our direct data. Derived data usually draws on other people's
aggregated data and has predictive powers going way beyond what we say in
our directly posted data.

Many policy-makers have also proposed compensating the public for the data
they give to online sites.

Companies usually guard derived data assiduously and refuse to delete it
when we ask them to delete our direct data. So I wonder whether free
software proponents should propose something like the GPK or copyleft for
user data: if you use my data, you have to give me access to whatever
insights you derive from it.

Has this been discussed anywhere?


Andy Oram  |  Editor
O'Reilly Media, Inc.  |  617-499-7479 |  oreilly.com

[image: oreilly_email_logo.png] <http://oreilly.com/>
_______________________________________________
libreplanet-discuss mailing list
libreplanet-discuss@libreplanet.org
https://lists.libreplanet.org/mailman/listinfo/libreplanet-discuss

Re: Direct data and derived data: chance for a data GPL?

[Taiidan]

[-- Attachment #1: Type: text/plain, Size: 944 bytes --]

I however do not want "insights" period.

How about something much - a law that says companies can't (ab)use the
data of paid users without their explicit opt-in and can't provide any
conditionals or anything more than minor incentives for opting in.

minor incentives - eg: $10/mo ok but $100 not ok.


More laws:
* The right to be forgotten where when one establishes an account at say
a bank, doctors office, store, airline etc they can simply check a box
that says "delete my information after X amount of time"

* Only the bare minimum of data required to provide a service should be
collected - there is NO reason a transport company needs your name to
transport you or a cable company your name to give you cable as long as
you pay for your cable modem and pre-pay for services.


All three should be a worldwide treaty - WILL be a worldwide treaty when
some day the celebrity leaks become relentless and the shiny people
demand change.

[-- Attachment #2: 0xDF372A17.asc --]
[-- Type: application/pgp-keys, Size: 5331 bytes --]

[-- Attachment #3: Type: text/plain, Size: 183 bytes --]

_______________________________________________
libreplanet-discuss mailing list
libreplanet-discuss@libreplanet.org
https://lists.libreplanet.org/mailman/listinfo/libreplanet-discuss

Re: Direct data and derived data: chance for a data GPL?

[Andy Oram]

Thanks, Taiidan. I think the European GDPR covers the two principles you
stated. (Compensation isn't covered.) In the US, a number of consumer
rights statements also include those two principles, but most companies
just do whatever they want.

Andy Oram  |  Editor
O'Reilly Media, Inc.  |  617-499-7479 |  oreilly.com

[image: oreilly_email_logo.png] <http://oreilly.com/>



On Mon, Aug 20, 2018 at 10:20 PM, Taiidan@gmx.com <Taiidan@gmx.com> wrote:

> I however do not want "insights" period.
>
> How about something much - a law that says companies can't (ab)use the
> data of paid users without their explicit opt-in and can't provide any
> conditionals or anything more than minor incentives for opting in.
>
> minor incentives - eg: $10/mo ok but $100 not ok.
>
>
> More laws:
> * The right to be forgotten where when one establishes an account at say
> a bank, doctors office, store, airline etc they can simply check a box
> that says "delete my information after X amount of time"
>
> * Only the bare minimum of data required to provide a service should be
> collected - there is NO reason a transport company needs your name to
> transport you or a cable company your name to give you cable as long as
> you pay for your cable modem and pre-pay for services.
>
>
> All three should be a worldwide treaty - WILL be a worldwide treaty when
> some day the celebrity leaks become relentless and the shiny people
> demand change.
>
> _______________________________________________
> libreplanet-discuss mailing list
> libreplanet-discuss@libreplanet.org
> https://lists.libreplanet.org/mailman/listinfo/libreplanet-discuss
>
_______________________________________________
libreplanet-discuss mailing list
libreplanet-discuss@libreplanet.org
https://lists.libreplanet.org/mailman/listinfo/libreplanet-discuss

Re: Direct data and derived data: chance for a data GPL?

[Ali Razeen]

And why should this be a law? How about leaving it to users to decide 
what they are okay with and what they are not? To use your analogy, if a 
transport company asks for a lot of information, and users are okay with 
that, who are you to say what they should or shouldn't do?

I am not a celebrity, nor someone famous nor rich. I really do not want 
a law like yours and I do *not* want it as a worldwide treaty. Please 
leave me my freedom, thanks.

-Ali


On 8/20/18 7:20 PM, Taiidan@gmx.com wrote:
> I however do not want "insights" period.
>
> How about something much - a law that says companies can't (ab)use the
> data of paid users without their explicit opt-in and can't provide any
> conditionals or anything more than minor incentives for opting in.
>
> minor incentives - eg: $10/mo ok but $100 not ok.
>
>
> More laws:
> * The right to be forgotten where when one establishes an account at say
> a bank, doctors office, store, airline etc they can simply check a box
> that says "delete my information after X amount of time"
>
> * Only the bare minimum of data required to provide a service should be
> collected - there is NO reason a transport company needs your name to
> transport you or a cable company your name to give you cable as long as
> you pay for your cable modem and pre-pay for services.
>
>
> All three should be a worldwide treaty - WILL be a worldwide treaty when
> some day the celebrity leaks become relentless and the shiny people
> demand change.
>
>
> _______________________________________________
> libreplanet-discuss mailing list
> libreplanet-discuss@libreplanet.org
> https://lists.libreplanet.org/mailman/listinfo/libreplanet-discuss


_______________________________________________
libreplanet-discuss mailing list
libreplanet-discuss@libreplanet.org
https://lists.libreplanet.org/mailman/listinfo/libreplanet-discuss

Re: Direct data and derived data: chance for a data GPL?

[Nicolás Ortega Froysa]

[-- Attachment #1.1: Type: text/plain, Size: 4467 bytes --]

I think you miss a certain point here. Firstly, he mentioned that these
laws could have an exception for when the company gives their users a
certain benefit from allowing them to do something with this information
of theirs. But also, what kind of people we are trying to protect with
these laws.

When it comes to social laws that are put in place, they're put in place
to protect the weaker members of society, people who perhaps don't
understand the full consequences of their actions. In terms of
technology, these people are the youth and the elderly. It makes sense
that a grown adult shouldn't be on a platform like Facebook if he's
concerned about his privacy, but neither a kid nor an elderly person
understands the ramifications of this. I have two grandparents who use a
lot of these technologies, not because they made the choice on their
own, but because they were told to by their children. They don't have an
education, so they don't understand anything, they just use it. There
are also kids who use the platform before they are even supposed to
based on the platform's own policy. In these cases, I would be in favor
of creating a law defending the rights of these individuals. One example
could be making it illegal to use the data of minors (while at the same
time not providing an incentive for them to give a false age, so the
company can mine their data).

It's also a matter of being given the choice. In many cases our choice
is "use the platform and lose your privacy, or don't use it". This
choice is very limiting, since in many cases we are forced to use a
platform for a variety of reasons (workplace, family pressure, only
means of communication in a given context, etc.). One should always be
given the option to use the platform without losing your privacy. I also
don't think that this option should require any more payment than the
other one, since this would be a form of income discrimination by which
the poorer users are more likely to sell off their privacy to use the
platform. Essentially, losing one's privacy should not receive discounts
in the price of the service.

On Mon, Aug 20, 2018 at 08:24:07PM -0700, Ali Razeen wrote:
> And why should this be a law? How about leaving it to users to decide what
> they are okay with and what they are not? To use your analogy, if a
> transport company asks for a lot of information, and users are okay with
> that, who are you to say what they should or shouldn't do?
> 
> I am not a celebrity, nor someone famous nor rich. I really do not want a
> law like yours and I do *not* want it as a worldwide treaty. Please leave me
> my freedom, thanks.
> 
> -Ali
> 
> 
> On 8/20/18 7:20 PM, Taiidan@gmx.com wrote:
> > I however do not want "insights" period.
> > 
> > How about something much - a law that says companies can't (ab)use the
> > data of paid users without their explicit opt-in and can't provide any
> > conditionals or anything more than minor incentives for opting in.
> > 
> > minor incentives - eg: $10/mo ok but $100 not ok.
> > 
> > 
> > More laws:
> > * The right to be forgotten where when one establishes an account at say
> > a bank, doctors office, store, airline etc they can simply check a box
> > that says "delete my information after X amount of time"
> > 
> > * Only the bare minimum of data required to provide a service should be
> > collected - there is NO reason a transport company needs your name to
> > transport you or a cable company your name to give you cable as long as
> > you pay for your cable modem and pre-pay for services.
> > 
> > 
> > All three should be a worldwide treaty - WILL be a worldwide treaty when
> > some day the celebrity leaks become relentless and the shiny people
> > demand change.
> > 
> > 
> > _______________________________________________
> > libreplanet-discuss mailing list
> > libreplanet-discuss@libreplanet.org
> > https://lists.libreplanet.org/mailman/listinfo/libreplanet-discuss
> 
> 
> _______________________________________________
> libreplanet-discuss mailing list
> libreplanet-discuss@libreplanet.org
> https://lists.libreplanet.org/mailman/listinfo/libreplanet-discuss

-- 
Nicolás Ortega Froysa
https://themusicinnoise.net/
http://uk7ewohr7xpjuaca.onion/
Public PGP Key:
https://themusicinnoise.net/nortega@themusicinnoise.net_pub.asc
http://uk7ewohr7xpjuaca.onion/nortega@themusicinnoise.net_pub.asc

[-- Attachment #1.2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

[-- Attachment #2: Type: text/plain, Size: 183 bytes --]

_______________________________________________
libreplanet-discuss mailing list
libreplanet-discuss@libreplanet.org
https://lists.libreplanet.org/mailman/listinfo/libreplanet-discuss

Re: Direct data and derived data: chance for a data GPL?

[bill-auger]

ah what a topic - so ideally suited for bike-shedding - i could do that
too but i will refrain and instead just answer with some cold facts

to address the OP's question directly, the GPL derives its authority
from copyright law; not from the author of the license, nor the author
of the work that uses the GPL - unless it can be shown that your tweet
to grandma saying: "mmm, i love these @nestle(TM) brand cupcakes" is a
copyright-able work of art then there are no rights that can be
retained from it, and therefore no copyleft-like principle can be
derived or enforced

furthermore, even if it were such that the data mined from these
so called "social" interactions (<190 characters? each) was
copyright-able (highly doubtful); that still would not imply that the
author has any rights to the thoughts that exists in the mind of the
reader - so there goes any claim to what emergent meta-relations or
correlations that have merely been "learned" or derived from a
collection of these things which were all voluntarily offered as public
knowledge, and offered for no purpose other than to become public
knowledge

that would be like publishing a book and then demanding that anyone who
weighs that book for the purpose of calculating the average weight of a
book, must share their result - swell idea, but i dont see where you
are going to attach it's feet nor it's teeth - the author has no
exclusive rights to the information content of the book's mass,
the person measuring it would not be breaking any existing law, and
there is no defensible moral objection to the activity either

lets say i ask every "social" website i ever ever used to delete me and
evey digital bit i have ever published to their service - yay - then
some unrelated third-party data-miner-bot "learns" through the public
APIs those services expose of the conspicuous holes i left in the
membership rosters and it "derives" a new correlation: that i am the
type of person who is likely to delete all my "social" accounts - so,
where exactly would i request that new "derived" information be deleted
from? - i would presumably not even know that this happened; but i
should be fully aware that anyone (or robot) with a computer can do it -
now lets say it was not a lone data-miner but 100 of them over the
course of several years; and they are not even interested in my
information anymore, because they had already milked it for everything
it was worth years ago - then they deleted it themselves after
exhausting it of value - yay, it got deleted, just like i wanted -
right? - clearly, the notion of data harvesting being restricted to an
opt-in/opt-out basis is ridiculously naive

some nice ugly facts for the skeptical:

1 ) the internet is public and decentralized

2 ) the information you post on the internet is mostly public

3 ) the information you post on the internet is mostly trivial

4 ) anything trivial is generally not copyright-able

5 ) the information that you choose to post publicly is widely available

6 ) people who are interested in any public information will get it

7 ) people who collect sets of similar information will collect it

8 ) people who get paid to analyze correlations
    across such similar collected data sets will do that too

9 ) people do get paid very well to do that too

10) all of the above is perfectly legal and actively encouraged
    by the very services that publish your data on your behalf

11) you are not required to use any of those services

12) you are not required to publish any data to the public internet

13) you are not required to use the internet

14) you are not required to use any computer

so, what else other than that are people actually complaining about? -
where exactly is the perceived problem there? - which one of those
points do people want to regulate or legislate? - are any of those
points incorrect? - did i omit something important?

let me give an example - i can say with 100% certainty that no one
is harvesting any information that i have posted on the 'twitter'
service - can anyone guess from which one of the above points that i can
conclude such certainty? - this is not rocket science, people

furthermore, being aware of #5 alone, allows me no excuse to complain
about where the words i type into the internet end up - once those
bits and bytes leave my machine, they are forever out of my hands; and i
really do not want to take them back, nor to prevent anyone else from
reading them - if that were the case then, i would not have typed them
in the first place

_______________________________________________
libreplanet-discuss mailing list
libreplanet-discuss@libreplanet.org
https://lists.libreplanet.org/mailman/listinfo/libreplanet-discuss

Re: Direct data and derived data: chance for a data GPL?

[Thomas Harding]

(Sorry for top-posting, due to handeld appliance)

Databases are "a set" of informations, and copyright applies on that collection as whole, and moreover on its structure.

Just take a look at OpenStreetMap and its licence.

Regarding personal data, it *should* be owned by individuals themselves on their own data, but thats not the case.

Just take a look at phone numbers databases, or mail adresses, that are a very expensive business.

And, regarding americans, anything reliable to insurance.

Best regards,
TSFH

Le 27 août 2018 15:33:20 GMT+02:00, bill-auger <bill-auger@peers.community> a écrit :
>ah what a topic - so ideally suited for bike-shedding - i could do that
>too but i will refrain and instead just answer with some cold facts
>
>to address the OP's question directly, the GPL derives its authority
>from copyright law; not from the author of the license, nor the author
>of the work that uses the GPL - unless it can be shown that your tweet
>to grandma saying: "mmm, i love these @nestle(TM) brand cupcakes" is a
>copyright-able work of art then there are no rights that can be
>retained from it, and therefore no copyleft-like principle can be
>derived or enforced
>
>furthermore, even if it were such that the data mined from these
>so called "social" interactions (<190 characters? each) was
>copyright-able (highly doubtful); that still would not imply that the
>author has any rights to the thoughts that exists in the mind of the
>reader - so there goes any claim to what emergent meta-relations or
>correlations that have merely been "learned" or derived from a
>collection of these things which were all voluntarily offered as public
>knowledge, and offered for no purpose other than to become public
>knowledge
>
>that would be like publishing a book and then demanding that anyone who
>weighs that book for the purpose of calculating the average weight of a
>book, must share their result - swell idea, but i dont see where you
>are going to attach it's feet nor it's teeth - the author has no
>exclusive rights to the information content of the book's mass,
>the person measuring it would not be breaking any existing law, and
>there is no defensible moral objection to the activity either
>
>lets say i ask every "social" website i ever ever used to delete me and
>evey digital bit i have ever published to their service - yay - then
>some unrelated third-party data-miner-bot "learns" through the public
>APIs those services expose of the conspicuous holes i left in the
>membership rosters and it "derives" a new correlation: that i am the
>type of person who is likely to delete all my "social" accounts - so,
>where exactly would i request that new "derived" information be deleted
>from? - i would presumably not even know that this happened; but i
>should be fully aware that anyone (or robot) with a computer can do it
>-
>now lets say it was not a lone data-miner but 100 of them over the
>course of several years; and they are not even interested in my
>information anymore, because they had already milked it for everything
>it was worth years ago - then they deleted it themselves after
>exhausting it of value - yay, it got deleted, just like i wanted -
>right? - clearly, the notion of data harvesting being restricted to an
>opt-in/opt-out basis is ridiculously naive
>
>some nice ugly facts for the skeptical:
>
>1 ) the internet is public and decentralized
>
>2 ) the information you post on the internet is mostly public
>
>3 ) the information you post on the internet is mostly trivial
>
>4 ) anything trivial is generally not copyright-able
>
>5 ) the information that you choose to post publicly is widely
>available
>
>6 ) people who are interested in any public information will get it
>
>7 ) people who collect sets of similar information will collect it
>
>8 ) people who get paid to analyze correlations
>    across such similar collected data sets will do that too
>
>9 ) people do get paid very well to do that too
>
>10) all of the above is perfectly legal and actively encouraged
>    by the very services that publish your data on your behalf
>
>11) you are not required to use any of those services
>
>12) you are not required to publish any data to the public internet
>
>13) you are not required to use the internet
>
>14) you are not required to use any computer
>
>so, what else other than that are people actually complaining about? -
>where exactly is the perceived problem there? - which one of those
>points do people want to regulate or legislate? - are any of those
>points incorrect? - did i omit something important?
>
>let me give an example - i can say with 100% certainty that no one
>is harvesting any information that i have posted on the 'twitter'
>service - can anyone guess from which one of the above points that i
>can
>conclude such certainty? - this is not rocket science, people
>
>furthermore, being aware of #5 alone, allows me no excuse to complain
>about where the words i type into the internet end up - once those
>bits and bytes leave my machine, they are forever out of my hands; and
>i
>really do not want to take them back, nor to prevent anyone else from
>reading them - if that were the case then, i would not have typed them
>in the first place
>
>_______________________________________________
>libreplanet-discuss mailing list
>libreplanet-discuss@libreplanet.org
>https://lists.libreplanet.org/mailman/listinfo/libreplanet-discuss

-- 
Je suis née pour partager, non la haine, mais l'amour.
Sophocle, /Antigone, 442 av. JC

_______________________________________________
libreplanet-discuss mailing list
libreplanet-discuss@libreplanet.org
https://lists.libreplanet.org/mailman/listinfo/libreplanet-discuss

Re: Direct data and derived data: chance for a data GPL?

[bill-auger]

On Tue, 28 Aug 2018 02:43:06 +0200 Thomas Harding <tom@thomas-harding.name> wrote:
> Databases are "a set" of informations, and copyright applies on that
> collection as whole, and moreover on its structure.

if thats true then it is a counter-argument to what the OP was
proposing - that siggest can only work if each individual user who
contributed each bit of data retains copyright on each bit - i dont
thik that is the case for such small bits as "tweets" and "likes"
whatnot - so if the copyright belongs to the collector then the users
have absolutely no grounds to ask them to delete or share any of it

_______________________________________________
libreplanet-discuss mailing list
libreplanet-discuss@libreplanet.org
https://lists.libreplanet.org/mailman/listinfo/libreplanet-discuss

Re: Direct data and derived data: chance for a data GPL?

[Aaron Wolf]

On 08/28/2018 05:37 AM, bill-auger wrote:
> On Tue, 28 Aug 2018 02:43:06 +0200 Thomas Harding <tom@thomas-harding.name> wrote:
>> Databases are "a set" of informations, and copyright applies on that
>> collection as whole, and moreover on its structure.
> 
> if thats true then it is a counter-argument to what the OP was
> proposing - that siggest can only work if each individual user who
> contributed each bit of data retains copyright on each bit - i dont
> thik that is the case for such small bits as "tweets" and "likes"
> whatnot - so if the copyright belongs to the collector then the users
> have absolutely no grounds to ask them to delete or share any of it
> 
> _______________________________________________
> libreplanet-discuss mailing list
> libreplanet-discuss@libreplanet.org
> https://lists.libreplanet.org/mailman/listinfo/libreplanet-discuss
> 

I'm not taking the time to fully engage on this topic, but note that
data / database issues are well-addressed by the dedicated community
around Open Knowledge.

So, among various interconnected sites: https://opendatacommons.org/

Their license *is* a copyleft GPL analogue for data.
https://opendatacommons.org/licenses/odbl/

_______________________________________________
libreplanet-discuss mailing list
libreplanet-discuss@libreplanet.org
https://lists.libreplanet.org/mailman/listinfo/libreplanet-discuss

Re: Direct data and derived data: chance for a data GPL?

[Thomas Harding]

There are 2 aspects

one adressing data copyright (and odbl works fine with copyright on data freeing).

The second on personal unanonymized data (health, wealth, political opinion, phone numbers, mail address, location, sex, etc), which should be an exception and should rely on a deletion mechanism (eg through a record "stamp" metadata comprising the origin database identifier and an individual identifier on that database, combined to a peer protocol --- cancel was not systematically honoured in usenet, but that's the idea).



Le 28 août 2018 19:49:32 GMT+02:00, Aaron Wolf <wolftune@riseup.net> a écrit :
>On 08/28/2018 05:37 AM, bill-auger wrote:
>> On Tue, 28 Aug 2018 02:43:06 +0200 Thomas Harding
><tom@thomas-harding.name> wrote:
>>> Databases are "a set" of informations, and copyright applies on that
>>> collection as whole, and moreover on its structure.
>> 
>> if thats true then it is a counter-argument to what the OP was
>> proposing - that siggest can only work if each individual user who
>> contributed each bit of data retains copyright on each bit - i dont
>> thik that is the case for such small bits as "tweets" and "likes"
>> whatnot - so if the copyright belongs to the collector then the users
>> have absolutely no grounds to ask them to delete or share any of it
>> 
>> _______________________________________________
>> libreplanet-discuss mailing list
>> libreplanet-discuss@libreplanet.org
>> https://lists.libreplanet.org/mailman/listinfo/libreplanet-discuss
>> 
>
>I'm not taking the time to fully engage on this topic, but note that
>data / database issues are well-addressed by the dedicated community
>around Open Knowledge.
>
>So, among various interconnected sites: https://opendatacommons.org/
>
>Their license *is* a copyleft GPL analogue for data.
>https://opendatacommons.org/licenses/odbl/
>
>_______________________________________________
>libreplanet-discuss mailing list
>libreplanet-discuss@libreplanet.org
>https://lists.libreplanet.org/mailman/listinfo/libreplanet-discuss

-- 
Je suis née pour partager, non la haine, mais l'amour.
Sophocle, /Antigone, 442 av. JC

_______________________________________________
libreplanet-discuss mailing list
libreplanet-discuss@libreplanet.org
https://lists.libreplanet.org/mailman/listinfo/libreplanet-discuss