Re: The Dangers of Web Annotation

Hi,

On the IA steering call yesterday, we had a good discussion on this, thanks in particular to Ivan and also to Conor Delahunty and Peg Fowler on our team that made good suggestions:

Here’s what we propose:

1- In order to preserve space for presentations, we’ll target one of the open afternoon slots for a session on “Consent and Abuse”— two quite differernt, but related, concepts.
2- We’ll do this on the first day, and most likely in the earlier part of the afternoon.
3- This will create 90 minutes that we can set aside for the issue.
4- We’ll attempt for the moment to provide space for treatment of both abuse and consent (treated either separately or together).  [The concepts of abuse/harrassment and consent are distinct, but related.  Consent (i.e. the question of whether page owners should be able to request that their content not be annotated, and how annotation platforms and services should react in kind) is the one that more specifically revolves around annotation per-se.  Abuse is a critical issue that annotation services will need to address in increasingly sophisticated ways, but it’s not exclusive to annotation.]
5- We’ll do some research to understand whether there are any folks that deal with reputation, abuse/trolls/spam, or the mechanics of the questions around consent that either a) live in Europe, or b) would be willing to join us (travel $ possible for 1-2 folks).
6- I will probably reach out to Genius to see if they’re open to joining a conversation about that.
7- I think there’s a balance between invited speakers who can provide background and context and a facilitated discussion.  We’ll try to strike that balance.  This is a subject many will likely feel like talking about, and I’m quite looking forward to the range of perspectives.
8- It’s possible that it may be difficult at this late date to get good folks to contribute. So we may need to improvise on format.

I need help.

- Specifically, suggestions (feel free to contact me personally) of either individuals, or contacts that might know individuals who would be good candidates to participate in such a session as presenters, panelists, etc.  These are folks w/ domain expertise.
- Suggestions as to someone who might be particularly good at facilitating such a session (preferably someone already in Europe, or who will be attending).
- Input into the design of a session.
- Input on how to treat these two concepts, individually or together.
- Any other thoughts.



On March 31, 2016 at 6:48:21 AM, Robert Sanderson (azaroth42@gmail.com) wrote:


All,

My 2c on the topic...

It would be great to surface the issue and discuss potential ways in which the situation can be improved.  I Annotate seems like a very good forum for the discussion, and I agree that having the right people there (whoever that might be) is important. 

That said, I also strongly agree with Ivan that the WG should not get derailed.  A solution to "Don't Read the Comments" requires many levels of technology above our simple model and protocol.  Dan hosted a fantastic workshop on reputation systems, with a focus on annotation, which underscored to me the range of potential areas that has a bearing on the topic, from identification of people (anonymity, pseudonymity, individual identity and global identity) to reputation networks and domains of expertise. Perhaps Dan might suggest folks from that workshop for I Annotate?

Anything we added at this stage to our specifications would be at best half baked and at worst dangerously wrong. We have pushed authentication and authorization off, even to the level of knowing whether the user is able to modify the annotation that they can see. Given that authentication seems, to me, like a necessary pre-requisite for even the most coarse solution of blocking users, let alone more complex and adaptive systems based around moderation and reputation, this will be a challenge for some time to come.

Rob

On Thu, Mar 31, 2016 at 3:32 AM, Ivan Herman <ivan@w3.org> wrote:
Hi Dan, Doug, everyone…

I think we all agree that I Annotate is a very timely occasion to have this on the agenda, and we have to do this. I do believe we should *try* to devote a full session (ie, 90 minutes) for this with, hopefully, an active engagement of the audience.

However... The problem with a panel is that it is probably a bit too late to get very influential people come over to Berlin for this. (For example, to get a fair perspective, having somebody from Genius on the panel might be a good thing, even if we may not agree with what they do, but I do not think it is realistic.) Maybe we should organize it in some very 'hackaton' like event; somebody (you Dan?) could introduce the topic in somewhat more details (I must admit that I do not understand all the intricacies either, for example), and we may just go on with a free form discussion and see where it goes. We could cherry pick some speakers from the speakers' list, but I am not sure we know in advance who would be the right persons.

I also have some more general question to myself (I do not expect ready made answers): I still want to understand what exactly the differences are between the harassments via annotation systems and harassments via, say, Facebook. We know that the latter has become a very serious problem, e.g., among teenagers, among political activists, around some religious communities; do we have something special here? Is it realistic, is it good, etc, to try to find a solution on the issue in isolation from that more general landscape? I really do not have an answer to these questions, and I would love to come out wiser from a discussion at I Annotate. 

My questions are also the reason why we should be careful not to put an extra responsibility on the the WA Working Group around this: I do not know whether the real problems, and their possible solutions, are related to the work the WA WG is doing or not. (Ie, whether the issue can be handled on the model/protocol level; my gut says the answer is no…). So the discussion should also concentrate to find out what the right forum is for this discussion, who should be around the table; I very much like Bill's comment:

"conversation should include those affected by it, and would strongly recommend reaching out *directly* to those who have been the subject of it. "

which does suggest that we may want to set up a separate, dedicated group on this. But, again, we may want to do that with experts in the social web area in general.

Sigh. I must admit, and I am not proud of this, that I never thought this issue would come to the fore around annotations. Yes, I am just naïve, or worse…

Ivan



On 30 Mar 2016, at 18:25, Dan Whaley <dwhaley@hypothes.is> wrote:

Sorry it’s taken me so long to respond. :

Yes, the last several days have been interesting.  It’s clear that there is a) a technological magic that sets annotation off from other kinds of discussion (twitter, reddit), and b) a certain lack of self-awareness on the part of some organizations at play here.

There is a narrative argued here (https://hyp.is/AVPIT_MqH9ZO4OKSldzh) that somehow annotations are an overlay that users bring to the page, and that this is users’ choice and page owners have no agency to protest.  While there are some obvious merits to that perspective, I think it ignores what we’re all actually trying to do here, which is to create something that ultimately ends up (more or less) in browsers, or at the very least as a pervasive technology that is widely used and generally visible.

We at H. believe that these issues must be addressed head on, and its up to all of us within the WAWG, and the larger open annotation community generally to do so.  It’s possible that there may be a role for standards here as well—curious as to others’ thoughts on that.  

Several themes are emerging:

a) The notion of “consent to be annotated", i.e. can folks opt-out of annotation on their pages, and what would an opt-out mean. i.e. is it a request that can be overridden if something is in the public interest?  How do we preserve the notion of “Truth to power” that annotation represents, while giving people that create (for instance) personal pages that say very personal things some protection. (Imagine that the government of Turkey sets their sites to all be opt-out, should we respect that? Would we respect the personal blog of an [ISSUE] skeptic whose posts are widely cited and used as evidence in larger debates?).  How would such an opt-out be signaled?  Is there a standard metatag on the page that indicates a preference of the page author?  an “annotate.txt” file?  etc.   What are the things one might opt out of?  (annotation, or only some kinds of annotation [you might still want to receive copy-edits], being proxied, etc.)

b) Approaches to deal with abuse.  What are the range of policies, automation and flagging (among other things) that annotation providers should employ to lead responsible communities.

c) What else?

I absolutely think that a session, probably a panel, or panel-assisted larger group discussion, at IA on this is the perfect thing.  
Maybe a double slot. That would mean reducing the number of presentations we can accept (happening separately by the program committee).  

Another option would be to swap in this session for one of the unconference sessions in the afternoon (perhaps a better solution).  

What I’d love suggestions on is:
- Format (panel, facilitated group discussion, hybrid, something else).
- Length
- When
- Who— if we do have some speakers, who might they be.  Are there folks in Europe people know who are particularly thoughtful about these issues?  Range of perspectives is helpful.  

Thanks for bringing this up Doug, and for the follow on discussion.

Dan

On March 30, 2016 at 7:58:03 AM, Salisbury, Davis - United States (dsalisbury@wiley.com) wrote:

Just weighing in to say that I believe that this is an important issue. Seems like a perfect thing to discuss at iAnnotate. I think the wider conversations around the Web are doing a good job of bringing issues to light and getting the conversation going, but it would be good for this group to have a stance and to contribute to it as well. Dan's blog post is a good start. 

Regards, 
Davis Salisbury 

-----Original Message----- 
From: Doug Schepers [mailto:schepers@w3.org]  
Sent: Wednesday, March 30, 2016 10:49 AM 
To: Siegman, Tzviya - Hoboken; Ivan Herman 
Cc: W3C Public Annotation List; Dan Whaley 
Subject: Re: The Dangers of Web Annotation 

Thanks for passing that on, Tzviya! 

As expected, Dan and the Hypothes.is team are thoughtful and responsive to the social implications of their work, and I appreciate their leadership here. 

A session at I Annotate would be useful, and the various conversations happening around the Web need to happen in their own spaces; I hope that we can add to the dialog by focusing some conversation on the technical and policy sides. 

Regards– 
Doug 

On 3/30/16 7:25 AM, Siegman, Tzviya - Hoboken wrote: 
> Dan blogged about it yesterday [1]. Great idea to discuss at iAnnotate. 
> 
> [1] https://hypothes.is/blog/preventing-abuse/ 
> 
> Tzviya Siegman 
> Digital Book Standards & Capabilities Lead Wiley 
> 201-748-6884 
> tsiegman@wiley.com 
> 
> -----Original Message----- 
> From: Ivan Herman [mailto:ivan@w3.org] 
> Sent: Wednesday, March 30, 2016 5:18 AM 
> To: Doug Schepers 
> Cc: W3C Public Annotation List; Dan Whaley 
> Subject: Re: The Dangers of Web Annotation 
> 
> (cc-ing Dan explicitly, to draw his attention:-) 
> 
> I wonder whether it is still possible to set up a panel on this issue at I Annotate. We do not have the program finalized yet, so a slot for this would be great. 
> 
> The only problem is: who would moderate and, even more importantly, who could/would participate on a panel like that? 
> 
> Ivan 
> 
> 
>> On 30 Mar 2016, at 10:50, Doug Schepers <schepers@w3.org> wrote: 
>> 
>> Hi, folks– 
>> 
>> We've focused mostly on 3 things in this group: 
>> 
>> 1) the annotation model 
>> 2) the annotation REST protocol 
>> 3) the anchoring mechanism (e.g. FindText API) 
>> 
>> The charter describes other deliverables [1]; my annotation architecture diagram goes into details on a few more [2]. 
>> 
>> But we haven't really discussed the social implications of Web Annotations, outside of some informal chats. Specifically, we haven't determined notification and curation models, which are critical if Web Annotations are to be used as a social good, rather than an avenue for harrassment; nor have we discussed the idea of opting-in or opting-out of allowing annotations a particular site. 
>> 
>> There's been an interesting (if disturbing) thread the past few days about how Genius is being used for what could be considered harassment (and for rude comments, at the very mildest). I suggest that we read and discuss the blog post [3], Medium articles [4], tweets [5 – 10], and Github issues [11] that describe this abuse, and try to think about what our role, as technologists and standards folks, can do to help the situation. 
>> 
>> Ultimately, if Web Annotation does take off as a feature of the Web, these cases will become all too common. And I don't think that scholarly and academic uses will be immune (though the accountability and reputation risk will reduce abuse). And if such abuse continues, it reduces the value and incentive for Web Annotation to succeed at all. 
>> 
>> I don't want to derail the current push towards Recommendation, but I do think it behooves us to treat this seriously, maybe on this list, or maybe in other forums, such as I Annotate, and to discuss it with the broader community on social media, where they have started the conversation. 
>> 
>> Thoughts? 
>> 
>> 
>> [1] https://www.w3.org/annotation/charter/ 
>> [2] http://www.w3.org/annotation/diagrams/annotation-architecture.svg 
>> [3] 
>> https://ellacydawson.wordpress.com/2016/03/25/how-news-genius-silence 
>> s 
>> -writers/ [4] 
>> https://medium.com/@glennf/citation-appropriation-and-fair-use-news-g 
>> e 
>> nius-picks-up-again-where-failures-left-off-d640719a82ab#.exsmdb2l1 
>> [5] https://twitter.com/brosandprose/status/713380185001836544 
>> [6] https://twitter.com/brosandprose/status/714668474904264706 
>> [7] https://twitter.com/TheFriskyFairy/status/714843748199493633 
>> [8] https://twitter.com/FeralHomemaking/status/714970696867319810 
>> [9] https://twitter.com/krues8dr/status/714999625090994176 
>> [10] https://twitter.com/krues8dr/status/714999849205305345 
>> [11] https://github.com/opengovfoundation/madison/issues/920 
>> 
>> Regards– 
>> Doug 
>> 
> 
> 
> ---- 
> Ivan Herman, W3C 
> Digital Publishing Lead 
> Home: http://www.w3.org/People/Ivan/ 
> mobile: +31-641044153 
> ORCID ID: http://orcid.org/0000-0003-0782-2704 
> 
> 
> 
> 

—
+1 650 274 7647 cell


----
Ivan Herman, W3C 
Digital Publishing Lead
Home: http://www.w3.org/People/Ivan/
mobile: +31-641044153
ORCID ID: http://orcid.org/0000-0003-0782-2704







--
Rob Sanderson
Information Standards Advocate
Digital Library Systems and Services
Stanford, CA 94305
—
+1 650 274 7647 cell