Barbara Page, a Wikipedia Visiting Scholar and Wikipedian in Residence at the University of Pittsburgh, has written a blog post[1] for Wikipediocracy about how the Amazon Echo's Alexa assistant reads out Wikipedia articles in response to queries. This includes queries that do not specifically ask for Wikipedia information.
What's the deal with the CC licence here?
To quote from the Creative Commons Attribution-ShareAlike 3.0 Unported Licence identified at the bottom of each English Wikipedia page,
If You Distribute, or Publicly Perform the Work or any Adaptations or Collections, *You must,* unless a request has been made pursuant to Section 4(a), *keep intact all copyright notices for the Work and provide, reasonable to the medium or means You are utilizing: (i) the name of the Original Author (or pseudonym, if applicable) if supplied, and/or if the Original Author and/or Licensor designate another party or parties (e.g., a sponsor institute, publishing entity, journal) for attribution ("Attribution Parties") in Licensor's copyright notice, terms of service or by other reasonable means, the name of such party or parties; (ii) the title of the Work if supplied; (iii) to the extent reasonably practicable, the URI, if any, that Licensor specifies to be associated with the Work, unless such URI does not refer to the copyright notice or licensing information for the Work*; and (iv) , consistent with Section 3(b), in the case of an Adaptation, a credit identifying the use of the Work in the Adaptation (e.g., "French translation of the Work by Original Author," or "Screenplay based on original Work by Original Author").
Some similar services preface their Wikipedia readings with "According to Wikipedia, ..." This is at least a minimum amount of attribution. While I am not a legal expert, I guess it could be construed as an attempt to comply with the "reasonable to the medium or means" passage above. It also tells the user where the information comes from, which is useful from the standpoint of transparency.
But the Amazon Echo appears to include no attribution whatsoever when providing Wikipedia-based answers. On the face of it, this would seem to violate the terms of the Creative Commons licence (as well as obscuring the origin of the information provided). Am I missing something?
Has this ever been the subject of discussions, agreements or understandings between Amazon and WMF?
Best, Andreas
This is an excellent point. Yes Amazon should definitely state that what they are reading is from Wikipedia.
James
On Wed, Jul 26, 2017 at 7:11 PM, Andreas Kolbe jayen466@gmail.com wrote:
Barbara Page, a Wikipedia Visiting Scholar and Wikipedian in Residence at the University of Pittsburgh, has written a blog post[1] for Wikipediocracy about how the Amazon Echo's Alexa assistant reads out Wikipedia articles in response to queries. This includes queries that do not specifically ask for Wikipedia information.
What's the deal with the CC licence here?
To quote from the Creative Commons Attribution-ShareAlike 3.0 Unported Licence identified at the bottom of each English Wikipedia page,
If You Distribute, or Publicly Perform the Work or any Adaptations or Collections, *You must,* unless a request has been made pursuant to Section 4(a), *keep intact all copyright notices for the Work and provide, reasonable to the medium or means You are utilizing: (i) the name of the Original Author (or pseudonym, if applicable) if supplied, and/or if the Original Author and/or Licensor designate another party or parties (e.g., a sponsor institute, publishing entity, journal) for attribution ("Attribution Parties") in Licensor's copyright notice, terms of service or by other reasonable means, the name of such party or parties; (ii) the title of the Work if supplied; (iii) to the extent reasonably practicable, the URI, if any, that Licensor specifies to be associated with the Work, unless such URI does not refer to the copyright notice or licensing information for the Work*; and (iv) , consistent with Section 3(b), in the case of an Adaptation, a credit identifying the use of the Work in the Adaptation (e.g., "French translation of the Work by Original Author," or "Screenplay based on original Work by Original Author").
Some similar services preface their Wikipedia readings with "According to Wikipedia, ..." This is at least a minimum amount of attribution. While I am not a legal expert, I guess it could be construed as an attempt to comply with the "reasonable to the medium or means" passage above. It also tells the user where the information comes from, which is useful from the standpoint of transparency.
But the Amazon Echo appears to include no attribution whatsoever when providing Wikipedia-based answers. On the face of it, this would seem to violate the terms of the Creative Commons licence (as well as obscuring the origin of the information provided). Am I missing something?
Has this ever been the subject of discussions, agreements or understandings between Amazon and WMF?
Best, Andreas
[1] http://wikipediocracy.com/2017/07/24/alexa/ _______________________________________________ Wikimedia-l mailing list, guidelines at: https://meta.wikimedia.org/ wiki/Mailing_lists/Guidelines and https://meta.wikimedia.org/ wiki/Wikimedia-l New messages to: Wikimedia-l@lists.wikimedia.org Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l, mailto:wikimedia-l-request@lists.wikimedia.org?subject=unsubscribe
Maybe some fact checking before getting all upset would be a good idea?
The blog post is a good story, but doesn't actually reflect how Alexa works wrt searching WIkipedia (I just quickly reconfigured one of mine to US English just to verify).
Simon
Simon, could you clarify?
Can you configure the device to give attribution?
On Thu, Jul 27, 2017 at 5:56 PM, Simon Poole simon@poole.ch wrote:
Maybe some fact checking before getting all upset would be a good idea?
The blog post is a good story, but doesn't actually reflect how Alexa works wrt searching WIkipedia (I just quickly reconfigured one of mine to US English just to verify).
Simon
Wikimedia-l mailing list, guidelines at: https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines and https://meta.wikimedia.org/wiki/Wikimedia-l New messages to: Wikimedia-l@lists.wikimedia.org Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l, mailto:wikimedia-l-request@lists.wikimedia.org?subject=unsubscribe
You need to explicitly ask for a Wikipedia article to get one (and it ends reading the summary with "More from wikipedia?" or something along such lines). That kind of renders the attribution issue moot.
If you don't explicitly ask for Wikipedia it will search with bing and simply read the result of that (which in general doesn't really work), which may naturally contain information from WP, but not curated by Amazon.
I just jumped on this because I was reading one of comparative tests between the Apple, Google, Microsoft and Amazon voice controlled assistant products a couple of days back, and one of the not so good sides of Alexa was answering general knowledge questions, which it competitors get around by using WP automatically. This matches with my experience of the device (in German you need to install a skill with a rather cumbersome UI to get direct access to WP so it is even worse in the ease of access to WP department).
Simon
Am 27.07.2017 um 13:26 schrieb John Mark Vandenberg:
Simon, could you clarify?
Can you configure the device to give attribution?
On Thu, Jul 27, 2017 at 5:56 PM, Simon Poole simon@poole.ch wrote:
Maybe some fact checking before getting all upset would be a good idea?
The blog post is a good story, but doesn't actually reflect how Alexa works wrt searching WIkipedia (I just quickly reconfigured one of mine to US English just to verify).
Simon
Wikimedia-l mailing list, guidelines at: https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines and https://meta.wikimedia.org/wiki/Wikimedia-l New messages to: Wikimedia-l@lists.wikimedia.org Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l, mailto:wikimedia-l-request@lists.wikimedia.org?subject=unsubscribe
Simon,
If you look at the comments under Barbara's piece, Greg linked to this YouTube video:
https://www.youtube.com/watch?v=ZurOYgQLT44
I had a look at that video before posting here. (I think it's kind of a daft video, but it does a perfectly good job of demonstrating how the Echo works.)
In this video, the lady asks at the beginning, "Alexa, who is Edward Snowden?"
The response reflects the lead sentence of the Wikipedia article, such as it was at the time.
At 0:30 in the video, she asks "Alexa, who is the FBI?" Again, Alexa responds with the lead sentence of Wikipedia's FBI article as it was at the time.
So she did not explicitly ask for the Wikipedia article, and yet got Wikipedia content. And I wonder, if she had said "Tell me more", would she have gotten more from Wikipedia at that point?
You say that Alexa reportedly gets some of this from Bing. But even if that's the case, how does it make a difference? To me it seems rather like Flickrwashing (Bingwashing?).
The end result is that the chain of attribution is broken and the content is treated as though it were public domain rather than published under a Creative Commons licence.
Andreas
On Thu, Jul 27, 2017 at 1:03 PM, Simon Poole simon@poole.ch wrote:
You need to explicitly ask for a Wikipedia article to get one (and it ends reading the summary with "More from wikipedia?" or something along such lines). That kind of renders the attribution issue moot.
If you don't explicitly ask for Wikipedia it will search with bing and simply read the result of that (which in general doesn't really work), which may naturally contain information from WP, but not curated by Amazon.
I just jumped on this because I was reading one of comparative tests between the Apple, Google, Microsoft and Amazon voice controlled assistant products a couple of days back, and one of the not so good sides of Alexa was answering general knowledge questions, which it competitors get around by using WP automatically. This matches with my experience of the device (in German you need to install a skill with a rather cumbersome UI to get direct access to WP so it is even worse in the ease of access to WP department).
Simon
Am 27.07.2017 um 13:26 schrieb John Mark Vandenberg:
Simon, could you clarify?
Can you configure the device to give attribution?
On Thu, Jul 27, 2017 at 5:56 PM, Simon Poole simon@poole.ch wrote:
Maybe some fact checking before getting all upset would be a good idea?
The blog post is a good story, but doesn't actually reflect how Alexa works wrt searching WIkipedia (I just quickly reconfigured one of mine to US English just to verify).
Simon
Wikimedia-l mailing list, guidelines at: https://meta.wikimedia.org/
wiki/Mailing_lists/Guidelines and https://meta.wikimedia.org/ wiki/Wikimedia-l
New messages to: Wikimedia-l@lists.wikimedia.org Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l,
mailto:wikimedia-l-request@lists.wikimedia.org?subject=unsubscribe
Wikimedia-l mailing list, guidelines at: https://meta.wikimedia.org/ wiki/Mailing_lists/Guidelines and https://meta.wikimedia.org/ wiki/Wikimedia-l New messages to: Wikimedia-l@lists.wikimedia.org Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l, mailto:wikimedia-l-request@lists.wikimedia.org?subject=unsubscribe
Hi,
On 27/07/2017 14:36, Andreas Kolbe wrote:
If you look at the comments under Barbara's piece, Greg linked to this YouTube video:
since I don't own an Amazon Echo, I will have to rely on the video.
I had a look at that video before posting here. (I think it's kind of a daft video, but it does a perfectly good job of demonstrating how the Echo works.)
In this video, the lady asks at the beginning, "Alexa, who is Edward Snowden?"
The response reflects the lead sentence of the Wikipedia article, such as it was at the time.
At 0:30 in the video, she asks "Alexa, who is the FBI?" Again, Alexa responds with the lead sentence of Wikipedia's FBI article as it was at the time.
The video was posted on March 9th, 2017.
This is the article about Edward Snowden as of March 6th, 2017: https://en.wikipedia.org/w/index.php?title=Edward_Snowden&oldid=76889960...
and this is the article about FBI as of March 7th,2017: https://en.wikipedia.org/w/index.php?title=Federal_Bureau_of_Investigation&a...
The response about Edward Snowden is not verbatim (I had a look also at some of the previous revisions, and the incipit did not change). It is very similar for sure and I can see some way to automatically get from the Wikipedia article incipt to the sentence spoken by Alexa. But at this point if you rephrase a sentence and use facts I don't think you are infringing copyright. It could be akin to close paraphrasing[1], but the quantity of text is limited.
The response about FBI instead is verbatim.
In both cases, they may be within the realm of the "right to quote"[2] (I am not sure this concept exists in US law per se) or "fair use".
You say that Alexa reportedly gets some of this from Bing. But even if that's the case, how does it make a difference? To me it seems rather like Flickrwashing (Bingwashing?).
It may totally make a difference. I am not a lawyer, but I think the question about the copyright status of search snippets and indexes for search engines has already been addressed by jurisprudence.
Simply put, the amount of text used changes the situation from "right to quote"/"fair use" to "derivative work".
Furthermore, to correctly cite Wikipedia, if snippets would not be considered under fair use/right to quote, they would need to also cite the license.
In this regard, compare the difference - http://imgur.com/gallery/3FQZS - between the snippets (both from Google and Bing), which do not have a license indication and the extensive portion of text which is displayed in the box in Bing which correctly indicates both the link to the original work and the license. Interestingly, in the case of the FBI, the box in Bing has less text and no indication of the license. It may be that they automatically decide that if they are going to show more than N words/characters then they do not treat the text as a quotation but as a derivative work and so they show the license.
I tried with another couple searches and this behavior seems consistent. If they shw a short chunk of text (~ 1 sentence), they do not provide the source and link to the license. If they show a big chunk of text (with a "+" sign) they do.
The Wikimedia Foundation could ask for a clarification to Amazon, but I suspect that the answer would not differ very much from above.
Cristian
[1]: https://en.wikipedia.org/wiki/Wikipedia:Close_paraphrasing [2]: https://en.wikipedia.org/wiki/Right_to_quote
We want these devices to read from Wikipedia. We just want attribution as appropriate. If they are already attributing when they go beyond fair use than all is good.
J
On Thu, Jul 27, 2017 at 8:14 AM, Cristian Consonni cristian@balist.es wrote:
Hi,
On 27/07/2017 14:36, Andreas Kolbe wrote:
If you look at the comments under Barbara's piece, Greg linked to this YouTube video:
since I don't own an Amazon Echo, I will have to rely on the video.
I had a look at that video before posting here. (I think it's kind of a daft video, but it does a perfectly good job of demonstrating how the
Echo
works.)
In this video, the lady asks at the beginning, "Alexa, who is Edward Snowden?"
The response reflects the lead sentence of the Wikipedia article, such as it was at the time.
At 0:30 in the video, she asks "Alexa, who is the FBI?" Again, Alexa responds with the lead sentence of Wikipedia's FBI article as it was at
the
time.
The video was posted on March 9th, 2017.
This is the article about Edward Snowden as of March 6th, 2017: https://en.wikipedia.org/w/index.php?title=Edward_Snowden&oldid=76889960...
and this is the article about FBI as of March 7th,2017: https://en.wikipedia.org/w/index.php?title=Federal_ Bureau_of_Investigation&oldid=769027291
The response about Edward Snowden is not verbatim (I had a look also at some of the previous revisions, and the incipit did not change). It is very similar for sure and I can see some way to automatically get from the Wikipedia article incipt to the sentence spoken by Alexa. But at this point if you rephrase a sentence and use facts I don't think you are infringing copyright. It could be akin to close paraphrasing[1], but the quantity of text is limited.
The response about FBI instead is verbatim.
In both cases, they may be within the realm of the "right to quote"[2] (I am not sure this concept exists in US law per se) or "fair use".
You say that Alexa reportedly gets some of this from Bing. But even if that's the case, how does it make a difference? To me it seems rather
like
Flickrwashing (Bingwashing?).
It may totally make a difference. I am not a lawyer, but I think the question about the copyright status of search snippets and indexes for search engines has already been addressed by jurisprudence.
Simply put, the amount of text used changes the situation from "right to quote"/"fair use" to "derivative work".
Furthermore, to correctly cite Wikipedia, if snippets would not be considered under fair use/right to quote, they would need to also cite the license.
In this regard, compare the difference - http://imgur.com/gallery/3FQZS
- between the snippets (both from Google and Bing), which do not have a
license indication and the extensive portion of text which is displayed in the box in Bing which correctly indicates both the link to the original work and the license. Interestingly, in the case of the FBI, the box in Bing has less text and no indication of the license. It may be that they automatically decide that if they are going to show more than N words/characters then they do not treat the text as a quotation but as a derivative work and so they show the license.
I tried with another couple searches and this behavior seems consistent. If they shw a short chunk of text (~ 1 sentence), they do not provide the source and link to the license. If they show a big chunk of text (with a "+" sign) they do.
The Wikimedia Foundation could ask for a clarification to Amazon, but I suspect that the answer would not differ very much from above.
Cristian
Wikimedia-l mailing list, guidelines at: https://meta.wikimedia.org/ wiki/Mailing_lists/Guidelines and https://meta.wikimedia.org/ wiki/Wikimedia-l New messages to: Wikimedia-l@lists.wikimedia.org Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l, mailto:wikimedia-l-request@lists.wikimedia.org?subject=unsubscribe
On 27/07/2017 16:46, James Heilman wrote:
We want these devices to read from Wikipedia. We just want attribution as appropriate. If they are already attributing when they go beyond fair use than all is good.
I agree, but the only way to know for sure is asking, I think.
Also, I am assuming that they already thought about this because in the end they need to balance what they are legally obliged to do with what they want to do to have a smooth interaction with the user.
Cristian
Hi Cristian,
In both cases, they may be within the realm of the "right to quote"[2]
(I am not sure this concept exists in US law per se) or "fair use".
I haven't checked the accuracy of the "Right to quote" Wikipedia article you referenced, but I will note here that it mentions the following requirements:
- the cited paragraphs are within a reasonable limit (varying from country to country), - clearly marked as quotations https://en.wikipedia.org/wiki/Quotations and fully referenced, - the resulting new work is not just a collection of quotations, but constitutes a fully original work in itself.
Assuming for the moment that the information in the Wikipedia article is sound, I don't think the first requirement is met: we are talking about potentially millions of words reproduced from Wikipedia without attribution. The second one is clearly not met either. The third one is debatable, but I would lean towards the opinion that the entire body of Alexa's output is not a "fully original work in itself" if its answers draw substantially on Wikipedia.
The Wikipedia article's lead further mentions that the making of quotations has to be "compatible with fair practice, and their extent does not exceed that justified by the purpose".
The Amazon Echo is a commercial product sold for the company's (considerable) profit. The information "Alexa" provides is a substantial part of the product's value to the customer. It's the main selling point.
This kind of commercial purpose is completely different from the making of quotations in the context of critical reviews, parodies and so on, which in no way attempt to *replace* the source quoted from, but instead exist to provide a sort of commentary on it.
My understanding is that this latter kind of purpose is the kind of purpose "fair use" is meant to support.
You say that Alexa reportedly gets some of this from Bing. But even if that's the case, how does it make a difference? To me it seems rather
like
Flickrwashing (Bingwashing?).
It may totally make a difference. I am not a lawyer, but I think the question about the copyright status of search snippets and indexes for search engines has already been addressed by jurisprudence.
Surely, the difference with search snippets is that such snippets, by the very nature of things, include a link to the original. That is their purpose. And this has a direct bearing on all of the CC licence requirements quoted earlier. In other words, by following the link provided in the search result, readers can view the original work the snippet is taken from, identify its title, its authors, its copyright status and the precise licence under which the material is published.
None of that is possible with Alexa. Alexa's use of Wikipedia, where things literally come out of a black box, with the listener uninformed as to their origin, is substantially different, and indeed competitive to Wikipedia itself.
Simply put, the amount of text used changes the situation from "right to quote"/"fair use" to "derivative work".
Derivative works, of course, have to be published under the same licence.
Furthermore, to correctly cite Wikipedia, if snippets would not be considered under fair use/right to quote, they would need to also cite the license.
See above – search snippets contain a link to the original. This might be considered "reasonable" in the context of a search engine, given space limitations, usability considerations etc.
To be sure, in the case of the Amazon Echo, I don't think it would be "reasonable" to expect "Alexa" to recite the complete text of the CC licence, along with the names of all contributors who had worked on the material included, etc. The medium is a material factor here. But Amazon appears to proceed on the assumption that it is "reasonable" to proceed *without any attribution whatsoever* in their delivery medium.
In this regard, compare the difference - http://imgur.com/gallery/3FQZS
- between the snippets (both from Google and Bing), which do not have a
license indication and the extensive portion of text which is displayed in the box in Bing which correctly indicates both the link to the original work and the license. Interestingly, in the case of the FBI, the box in Bing has less text and no indication of the license. It may be that they automatically decide that if they are going to show more than N words/characters then they do not treat the text as a quotation but as a derivative work and so they show the license.
I tried with another couple searches and this behavior seems consistent. If they shw a short chunk of text (~ 1 sentence), they do not provide the source and link to the license. If they show a big chunk of text (with a "+" sign) they do.
Interesting, thank you.
The Wikimedia Foundation could ask for a clarification to Amazon, but I suspect that the answer would not differ very much from above.
A clarification and examination of Amazon's reasoning would indeed be welcome.
Andreas
Cristian
Wikimedia-l mailing list, guidelines at: https://meta.wikimedia.org/ wiki/Mailing_lists/Guidelines and https://meta.wikimedia.org/ wiki/Wikimedia-l New messages to: Wikimedia-l@lists.wikimedia.org Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l, mailto:wikimedia-l-request@lists.wikimedia.org?subject=unsubscribe
On Thu, Jul 27, 2017 at 3:14 PM, Cristian Consonni cristian@balist.es wrote:
The video was posted on March 9th, 2017.
This is the article about Edward Snowden as of March 6th, 2017: https://en.wikipedia.org/w/index.php?title=Edward_Snowden&oldid=76889960...
The response about Edward Snowden is not verbatim (I had a look also at some of the previous revisions, and the incipit did not change).
If you go back a bit further in the article's history, the odd "Ed" that Alexa reads out in the video, between "Joseph" and "Snowden", makes an appearance in Wikipedia as well. E.g. here:
https://en.wikipedia.org/w/index.php?title=Edward_Snowden&oldid=73686894...
Edward Joseph "Ed" Snowden ...
I will not spend an hour trying to identify the exact article version that matches Alexa's output in that video best, but it's safe to assume that this inserted "Ed", too, came from Wikipedia, even though it had gone by the time the video was uploaded to YouTube.
Am 27.07.2017 um 18:37 schrieb Andreas Kolbe:
Edward Joseph "Ed" Snowden ...
I will not spend an hour trying to identify the exact article version that matches Alexa's output in that video best, but it's safe to assume that this inserted "Ed", too, came from Wikipedia, even though it had gone by the time the video was uploaded to YouTube.
The current (full) answer is
'Edward Joseph "Ed" Snowden, the American computer professional former CIA employee, and government contractor who leaked classified information from the U.S. National Security Agency in 2013.'
Now obviously there could be -lots- going on behind the scenes, for example long term caching of search results (difficult to believe that Bing would allow that if it is really from them, but who knows) and so on.
Simon
On Thu, Jul 27, 2017 at 6:13 PM, Simon Poole simon@poole.ch wrote:
The current (full) answer is
'Edward Joseph "Ed" Snowden, the American computer professional former CIA employee, and government contractor who leaked classified information from the U.S. National Security Agency in 2013.'
What happens when you say "Tell me more"? Could you try please?
Now obviously there could be -lots- going on behind the scenes, for example long term caching of search results (difficult to believe that Bing would allow that if it is really from them, but who knows) and so on.
Yes, that's what I thought.
On Fri, Jul 28, 2017 at 1:29 AM, Adele Vrana avrana@wikimedia.org wrote:
Hello,
I am Adele Vrana, Director of Strategic Partnerships at the Foundation.
We have contacts at Amazon and will seek to clarify the questions raised on this thread. I will make sure to circle back with you once we have an update.
All the best, Adele
Thanks, Adele!
Some further thoughts on this thread while we wait for Adele to come back to us.
According to Statista,[1] the worldwide market for virtual digital assistants is expected to grow from $1.64 billion in 2015 to $15.79 billion by 2021. That's a tenfold increase over six years. Digital assistants will reportedly outnumber people by 2021.[2]
The Foundation has long been worried about the shift to mobile, the related decline in desktop page views, and its effect on donations. But doesn't the growing popularity of digital assistants – all of them, ironically, drawing to varying degrees on Wikimedia projects – represent a much greater threat to Wikipedia page views in the medium to long term than mobile?
At some point, I fear, opening a browser window on your computer or phone, typing in a search term, and opening up a Wikipedia page will to many people and in many contexts seem quaint and old-fashioned. Or to be more precise, it won't even seem quaint and old-fashioned: people will gradually stop doing it, and forget they ever did it, just like it's become rare for most of us to get up, walk to a bookcase, and look up a term in a book, because Googling is faster and more convenient. What will happen to Wikipedia page views then?
In my view, the decision taken many years ago to make Wikimedia content freely available to all re-users, even those earning billions of dollars from it, was a gigantic mistake. It sold contributors, who work for nothing, out. There should have been a (high) profit threshold above which re-users should have been required to pay a percentage of their profits back into the Wikimedia movement's pot.
But given that that bridge was burnt long ago – irretrievably so, it seems – shouldn't the Foundation now, at the very least, do its utmost to make sure that the minimal rights volunteers retain over the content they contribute to their projects are respected? (Wikidata's CC-0 licence was another huge mis-step in my view, as it reduces Wikimedia's visibility to nothing, just as it is reduced to nothing in the Amazon Echo examples described in this thread.) Otherwise, aren't we running the risk that Wikipedia may well be as healthy in a couple of decades' time as DMOZ is today?
And, bearing in mind the projected growth of the virtual digital assistant market, shouldn't the Wikimedia Foundation look at developing and marketing its own Wikipedia-based virtual digital assistant, to at least earn a share of the income its volunteers' work will generate in the years to come? Won't volunteers otherwise just continue to be fodder to make Apple, Microsoft, Google, Amazon etc. even richer than they already are?
Andreas
[1] https://www.statista.com/statistics/589079/worldwide- virtual-digital-assistants-consumer-market/ [2] http://www.netimperative.com/2017/05/digital- assistants-outnumber-people-2021/
There appears to already be an open-source AI voice assistant effort. https://www.forbes.com/sites/matthunckler/2017/05/15/this-open-source-ai-voi...
Like with the open search movement, IMO we should collaborate with the ongoing efforts being lead by others rather than lead the efforts ourselves or lead the efforts internally.
Agree we need to make sure that Wikipedia content is appropriately attributed by the wider world who re-use us. Our main goal IMO needs to be to make Wikipedia high quality / maintain its high quality.
James
On Mon, Jul 31, 2017 at 9:43 AM, Andreas Kolbe jayen466@gmail.com wrote:
Some further thoughts on this thread while we wait for Adele to come back to us.
According to Statista,[1] the worldwide market for virtual digital assistants is expected to grow from $1.64 billion in 2015 to $15.79 billion by 2021. That's a tenfold increase over six years. Digital assistants will reportedly outnumber people by 2021.[2]
The Foundation has long been worried about the shift to mobile, the related decline in desktop page views, and its effect on donations. But doesn't the growing popularity of digital assistants – all of them, ironically, drawing to varying degrees on Wikimedia projects – represent a much greater threat to Wikipedia page views in the medium to long term than mobile?
At some point, I fear, opening a browser window on your computer or phone, typing in a search term, and opening up a Wikipedia page will to many people and in many contexts seem quaint and old-fashioned. Or to be more precise, it won't even seem quaint and old-fashioned: people will gradually stop doing it, and forget they ever did it, just like it's become rare for most of us to get up, walk to a bookcase, and look up a term in a book, because Googling is faster and more convenient. What will happen to Wikipedia page views then?
In my view, the decision taken many years ago to make Wikimedia content freely available to all re-users, even those earning billions of dollars from it, was a gigantic mistake. It sold contributors, who work for nothing, out. There should have been a (high) profit threshold above which re-users should have been required to pay a percentage of their profits back into the Wikimedia movement's pot.
But given that that bridge was burnt long ago – irretrievably so, it seems – shouldn't the Foundation now, at the very least, do its utmost to make sure that the minimal rights volunteers retain over the content they contribute to their projects are respected? (Wikidata's CC-0 licence was another huge mis-step in my view, as it reduces Wikimedia's visibility to nothing, just as it is reduced to nothing in the Amazon Echo examples described in this thread.) Otherwise, aren't we running the risk that Wikipedia may well be as healthy in a couple of decades' time as DMOZ is today?
And, bearing in mind the projected growth of the virtual digital assistant market, shouldn't the Wikimedia Foundation look at developing and marketing its own Wikipedia-based virtual digital assistant, to at least earn a share of the income its volunteers' work will generate in the years to come? Won't volunteers otherwise just continue to be fodder to make Apple, Microsoft, Google, Amazon etc. even richer than they already are?
Andreas
[1] https://www.statista.com/statistics/589079/worldwide- virtual-digital-assistants-consumer-market/ [2] http://www.netimperative.com/2017/05/digital- assistants-outnumber-people-2021/ _______________________________________________ Wikimedia-l mailing list, guidelines at: https://meta.wikimedia.org/ wiki/Mailing_lists/Guidelines and https://meta.wikimedia.org/ wiki/Wikimedia-l New messages to: Wikimedia-l@lists.wikimedia.org Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l, mailto:wikimedia-l-request@lists.wikimedia.org?subject=unsubscribe
Am 27.07.2017 um 14:36 schrieb Andreas Kolbe:
... You say that Alexa reportedly gets some of this from Bing. But even if that's the case, how does it make a difference? To me it seems rather like Flickrwashing (Bingwashing?).
Christian already beat me to the punch and essentially has touched on all the in and outs of search results snippets and their copyright status, and I would defer to that and the WMF legal department for opinions on those aspects.
Wrt using bing, that is not "reportedly", one of the nice things about Alexa is that the companion app will actually display how it resolved the action. In both mentioned cases it uses a bing query to get an answer. The interesting bit is that the answer for Edward Snowden doesn't seem to turn up in the regular top bing search results even after a lot of reconfiguration.
In any case I don't quite understand the negative view on this, I would see it more as an opportunity for the WMF to improve how Amazon gets results from WP and try to cut out the middle man in this case.
Simon
Am 27.07.2017 um 16:58 schrieb Simon Poole:
Wrt using bing, that is not "reportedly", one of the nice things about Alexa is that the companion app will actually display how it resolved the action. In both mentioned cases it uses a bing query to get an answer. The interesting bit is that the answer for Edward Snowden doesn't seem to turn up in the regular top bing search results even after a lot of reconfiguration.
I may actually be jumping to conclusions on that, and maybe Amazon is not using bing for the simple search results in the end. But I still don't see a qualitative difference to a normal search result with all the caveats wrt copyright that Christian has mentioned.
Simon
wikimedia-l@lists.wikimedia.org