As more people share more content across social networks, the issue of who owns that content becomes increasingly important. Find out how the business practices of social networks mixed with pending (and current) global legislation could create the “perfect storm” of copyright-free chaos.
Orphaned Works Copyright Webinar
Social media networks are stripping embedded metadata, which could leave you without copyright ownership over things you share.
RECORDING DATE May 8, 2013
GUESTS Michael Steidl (IPTC); Jeff Sedlik (PLUS Coalition); David Riecks (ControlledVocabulary.com)
HOST David Diamond
Want to continue the discussion? Visit the YouTube page for this digital asset management webinar and leave a comment. If you wish to reach Mr. Riecks, Mr. Sedlik or Mr. Steidl, please email email@example.com and your request will be forwarded.
The resources below were mentioned during the webinar.
- International Press Telecommunications Council (IPTC) – iptc.org
- PLUS Coalition – PLUSCoalition.org
- PLUS Image Registry –PLUSregistry.org
- Controlled Vocabulary / Taxonomy Info and Resources – ControlledVocabulary.com
- Photo Metadata –PhotoMetadata.org
- CEPIC/IPTC Image Metadata Handbook »
- IPTC Photo Metadata User Guide » (PDF)
- Embedded Metadata Manifesto »
- IPTC Social Media Metadata Test Results »
- IPTC Photo Metadata Resources »
- 6 Ways to Keep your Images from Becoming “Orphan Works” »
Webinar Questions and Answers
A number of questions came in during the webinar that could not be answered due to time constraints. Those questions, along with answers from our panelists, can be found below.
What pressure can be brought? What’s the path forward? I know ASMP is on top of this, as is APA and others. How to make change? It’s been hard enough to get photographers to even add metadata. Now, we need follow through.
(Jeff Sedlik) Whatever your views on orphan works legislation, express them respectfully to your trade associations and your government representatives. Make an effort to learn about and understand the underlying challenges. Rather than demonizing those with opposing views, carefully consider those views and attempt to turn opponents into collaborators, seeking a solution that will solve the underlying problems and satisfy all concerned, to the greatest extent possible. Be prepared and willing to compromise. The hard line, no-holds-barred approach will most certainly fail. In speaking or writing publicly on the issue, acknowledge the underlying issues and indicate a willingness to seek a mutually beneficial solution. Make a special effort to avoid infighting in the ranks of your peers who might propose different strategies or approaches to solving the issue. Infighting will rapidly escalate and result in factions wasting their time and energy against each other, rather than making progress toward shared goals. Agree to disagree on certain points, and seek out points on which you can agree. Remember that if it is easy to find and contact you for permission to use your work, your work cannot be orphaned.
What are the possibilities to lock certain metadata to prevent tampering?
(David Riecks) I fully understand why photographers want their metadata to be permanent—they want it to be unalterable and perpetually associated with the image, regardless of the workflow through which it may travel. Unfortunately, today, unprotected data containers such as EXIF, IPTC/IIM, or XMP are inherently modifiable. It will take a significantly more complex system to change that, and that can’t happen overnight. Take a look at how long it has taken to get XMP support up to the level it is today (and that system still has a number of limitations); then add an additional component—like permanence—and you raise a number of conflicting requirements, making it exponentially harder to carry out.
The reality is that “protecting” metadata this way will require major architectural changes throughout all image editing, metadata annotation, and image database tools. At minimum, guaranteeing permanence in metadata would require new extensions to existing file formats which would also “break” backward compatibility. The applications, file formats, and workflows, which we currently use today are simply not designed for the more rigorous practices that will be needed to assure metadata and image permanence. As a result, insisting on permanent metadata would likely eliminate many legacy and open source (or small developer) tools from being used in an image workflow where metadata permanence is involved. Of course, so long as any application exists that can open a given file format and then resave it as another where metadata and image permanence are no longer guaranteed, it’s a moot point
(Michael Steidl) The short answer which covers all technical facets is no, the deletion or modification of metadata cannot be stopped by simple means. The only realistic option is to apply a secure certificate to the metadata. But this would cost some money and is not very simple to administer. Further, it raises the issue of overly sealed metadata (e.g. adding keywords would be impossible), and if the metadata headers as such are removed, the certificates don’t help, they only raise a warning if the metadata was modified.
(Jeff Sedlik) Although the the best current option is to embed your detailed metadata in your image files, the reality is that image files are fragile and inefficient vessels, ill suited for storing and communicating critical image metadata. Embedded Metadata can be modified or removed without your permission. Confidentiality concerns are also a significant issue limiting the use of embedded metadata to communicate detailed information that might be private or sensitive in nature. Also, image metadata is dynamic—it changes over time; but it is not possible to edit embedded image metadata after distribution of a file to bring that information current. The optimal method for communicating image metadata is to store an ID in your image file and provide a link to remotely stored, secure metadata that is controlled by you. Back that ID up as an invisible digital watermark in the file to allow recovery of the ID in the event that the embedded metadata is lost. Going forward, image recognition will be an important tool for image metadata recovery, but image recognition alone is an inadequate substitute for unique identifiers.
(Jeff Sedlik) Digimarc’s technology has dramatically improved and is worth a second look. The watermark is now is non-destructive and far more robust than earlier versions.
What can we do to get mobile phone support of metadata?
(David Riecks) Marksta allows you to add embedded metadata (as well as a visible watermark), but it’s for iOS only. Not aware of any comparable program yet for Android.
(Michael Steidl) My view is that metadata relevant for rights should be supported from the very start of a workflow. This means support by the operating system or by the camera app is required. Having to go through an additional step like opening the image and editing metadata with another app should be only an additional step.
Has Jeff invited Facebook, Twitter or others to join PLUS?
(Jeff Sedlik) Invited, but not yet members.
What is the benefit of stripping metadata from images for social media websites? It doesn’t seem like those metadata files would be very large.
(David Riecks) The usual arguments are that this shortens processing time (images are usually resized), and lowers the storage requirement. Small files will download quicker, though the difference to users is often negligible.
(Jeff Sedlik) I will first say that metadata should not be stripped without the permission of the person who added the metadata to file. That said, I’ll provide some background so that you can understand the factors at play. If you think that image metadata doesn’t take up much space, you’re not considering the issue at scale. For search engines and large sites dealing with many millions of images, the metadata in all of those files represents a very significant cost, both financially (in terms of moving that data around, serving it up repeatedly, backing it up) and in terms of performance. For a business model focused on generating impressions and click-throughs (e.g. Google), there is a focus on achieving the fastest possible page loads in order to achieve the greatest number of impressions and clicks in order to generate the greatest amount of revenue and maximize profits. Most pay attention to fine tuning the coding of pages to shave milliseconds off the time required to display each page. We took a file that was 64kb with minimal metadata, and we added metadata to all of the IPTC and PLUS fields. The resulting file size was 180kb, nearly 3 times larger. We then took a file that was 1×1 pixel, just 305 bytes, and added IPTC and PLUS metadata. The resulting file size with a full complement of metadata was 71 kilobytes (71,000 bytes), or 234 times larger. Of course, few people would populate *all* of the IPTC and PLUS metadata fields, and as for the second example, nobody is distributing 1×1 pixel images; but these examples serve to demonstrate two facts: (1) image metadata adds weight to a file, and (2) the smaller the image file size, the greater the effect of metadata on file size, proportional to the image file size. Again, this is not a viable justification for wiping metadata, but these are very real issues for companies dealing with image file metadata at scale, and if we don’t consider and deal with these issues, the metadata stripping issue will continue. The answer, again, is to embed an ID into the file, and store metadata externally. This results in an efficient file size, secure metadata, dynamic metadata, with access control, and removes any excuse for wiping the metadata to decrease file size.
What if there was an export feature, say in Lightroom, that would register an image with PLUS and the copyright office?
(Jeff Sedlik) PLUS has been working with the US Copyright Office for several years on this issue. We are currently waiting for the Copyright Office to provide a means by which external systems may connect to submit registrations. Once that is available, we will implement a copyright registration wizard to allow for registration of works with the copyright office when registering works with the PLUS Registry. PLUS is a global organization and the US is just one of 96 countries involved to date. Few countries provide a copyright registration option, but we are glad to work with the copyright office in any country offering an API for submission of copyright registrations. Submission of PLUS registrations from Lightroom, Aperture, Adobe applications and other applications is fundamental to the PLUS Registry API design. For clarification, the PLUS Registry is not a copyright registration system, it is a system connecting images to rights holders and rights information. This is something that copyright registration does not and will not provide. With few exceptions, copyright registration does not allow people to find rights holders for images. In fact, it is not possible to view images in the Copyright Office records unless you are the registrant or are involved in litigation arising from the work in question. Where the rubber meets the road, the primary benefit of copyright registration is the availability of an award of attorneys fees/costs and statutory damages, in that order, only if the registration is timely. This not only provides benefit in litigation, but also provides significant benefit in settlement discussions in order to avoid litigation.
When will we be able to start uploading images to the PLUS Registry?
(Jeff Sedlik) We are testing image registration and image search now, internally. If you have registered at www.PLUSregistry.org you will be notified when we open these functions to public beta testing. We are working as fast as we can, with the resources at hand. The development and operation of the Registry is funded by its users. Please consider upgrading to “Supporting Membership” by making a supporting contribution to support development.
What do the panelists think of StegMark?
(David Riecks) I’ve not used, so no comment. Would need to test first, before I could give any meaningful reply.
Regarding the image recognition software on PLUS. How effective is it at differentiating versions of an image, e.g. black and white images watermarked with a nondescript name such as John Smith?
(Michael Steidl) PLUS uses the PicScout technology. At a time when they [PicScout] were not part of Getty Images, they had on their website a showcase of variations of an image which could be identified as derivative work of the original. This was quite impressive: b/w versions, croppings (I guess ~50% of the area had to be left), versions with foggy overlays etc.
(Jeff Sedlik) In our tests, we found PicScout technology to be effective even when an image is modified (grayscaled, contrast, color, stripped out of background, combined with other images, etc). Image recognition alone is not the answer to the challenges facing our industry. The orphan works issue is only a symptom of a greater problem. We need to provide access to image metadata for the purposes of automated transactions and automated management of large quantities of images. Simply providing a link to contact information (email address, website, phone) for a rights holder might have been a suitable answer in the 90s, but not today, and not tomorrow. Photographers who believe that this is sufficient need to gain an understanding of the greater problem facing our industry and our ability to connect to the marketplace in an efficient manner. Also our ability to provide our clients with a means of managing our rights and avoiding unintentional infringement while dealing with huge quantities of images. The answer, again, is IDs in files, linked to remotely stored, machine interpretable image metadata.
Twitpic was in the media a while back for entering into an agreement with an entertainment news agency to sell images. The problem with this is that it creates an environment of mistrust between providers and users.
(David Riecks) I would agree, which is why it’s important to always read the “TOS” (Terms of Service) of any service, before joining or using.
One issue that hasn’t been touched on is the abuse of metadata by mainstream media. I sort of expect newer media like social media to strip metadata because in some cases they haven’t even thought about the issue. But that excuse doesn’t apply to mainstream media: their entire businesses are built on intellectual property and they’re fully aware of copyright law. Yet the BBC—the largest broadcasting organisation in the world—has an internal policy of stripping metadata from all photos: employees are instructed to do this. I recently had to analyse images on the website of a major US publisher that claimed they don’t strip metadata. Of 100 photos, 65% had some copyright information removed; 25% had all their metadata removed, completely orphaning the photos. It seems to me that so long as high profile professional media organisations continue to behave in such a fashion, it’s going to be difficult to persuade social media to do the right thing.
(David Riecks) When I last checked, BBC was retaining a minimal set of metadata. However, for many news outlets this is a problem. For many it’s an issue of their underlying CMS not supporting the preservation of metadata. If you see a problem, be sure to let the media company know.
I think we need to do two things as far as metadata is concerned. Firstly, campaign all these social networking sites to stop stripping out metadata. (In some territories, I believe the law makes this illegal anyway.) Campaign would involve distributing email addresses to organisations and associations and if need be, publicly, for campaigning purposes. Secondly, it will be to come up with a file format that locks in metadata, similar to a locked PDF document, and does not allow deletion or editing of metadata unless the password is shared.
(David Riecks) I’d say that whilst data storage and internet speed was initially the reason for stripping metadata, with the cost of storage and speed of broadband, I think that stripping metadata is no longer necessary and gives a notion of possibly sinister intentions.
(Jeff Sedlik) That would be wonderful but will not happen as long as legacy file types that allow image metadata to be removed remain in use. The various solutions we’re seeing that require storage of a parent copy of an image at one location, with all copies distributed using a link back to the parent copy; these solutions will fail because of copying that occurs once the images are distributed.
The notion that stripping metadata is not being done (by some parties) to add to plausible deniability under DMCA and related regs actually made me laugh out loud. Is this the “nice” webinar and there’s a secret one where we tell the truth?
(Jeff Sedlik) That is a statement, not a question, and it is a poorly worded statement at that. And it mischaracterizes the statements of the panelists. I didn’t hear anyone state that the stripping of metadata is “not being done.” In the other possible interpretation of your statement, I didn’t hear anyone state that the parties stripping metadata are not doing so purposefully. I did not hear any of us rule out any specific reason that a company might strip metadata. We discussed at least one reason—to increase the speed of page loads so as to generate more revenue from advertising exposures—but we did not rule out or reject any other reason. I did say that at the present time, in the US and in most countries, stripping metadata is not illegal unless it is done with the intent to infringe or to support infringement. That is a general statement, and the devil is in the details. There is a body of case law allowing search engines to display images for navigation purposes without the permission of rights holders. Standing by for litigation in the social media arena, and we’ll see what develops.
Jeff, you mention the embedded metadata and image recognition as routes to find the registered creator. But you didn’t mention steganographic methods, such as Digimark. Do you no longer feel that this method offers sufficient protection?
(Jeff Sedlik) Answered above.
Is there an integration of Picturepark DAM with the PLUS Registry?
(David Diamond [Picturepark moderator]) If it were up to me, it would be available right now. Picturepark prospects and customers are encouraged to submit feature requests at http://picturepark.com/requests
Each image needs an invisible barcode hidden in the file.
(David Riecks) This sounds more like a statement than a question, but this is one option that a registry can provide.
(Michael Steidl) There have been attempts of this kind of steganography, like reserving the least significant bit of pixel bytes for this purpose. But it was soon found that this does not work because as soon as the pixel content of an image is changed, the steganographic message is gone. From my experience the only way to identify an image is to create an algorithm which extracts the key visual characteristics of an image. PicScout and some other companies have developed such an algorithm.
(Jeff Sedlik) “Barcode” is not the correct concept and would not be practical or effective with images, but if you are referring to unique persistent identifiers pointing to remotely stored metadata, by all means yes, that is the answer. See www.PLUSregistry.org
How many different image ‘registries’ do you think the world can handle?
(Jeff Sedlik) As many as might be created. However, they must be linked together, so that a search of each registries searches all other registries, or rights holders and image users will be faced with an impossible challenge. Imagine a world with multiple registries in each country, with no connection between those registries. Sounds like a nightmare to me. In addition, the “hub” linking all of the registries together should be non-profit, controlled by all industries, and owned and operated by its users on a cost recovery basis. That is exactly what we’re doing at PLUS. Providing not just a Registry, but a hub to connect all registries, worldwide, operated under the control of all industries engaged in creating, distributing, using and preserving images.
You’ve hit the nail on the head! You need to put an ID into the file that cannot be removed that links to the metadata. You need the camera manufacturers onboard to do this.
(Michael Steidl) IPTC is aware that the camera makers have an important role in this area. I would like to extend this to smart phones with good to excellent built-in cameras. And we try to get in touch with them at our Photo Metadata Conference, 13 June in Barcelona (Spain) But it turns out to be hard because the makers are not very responsive on this topic.
(Jeff Sedlik) Yes, cooperation from the camera manufacturers would be a positive development. As images are and will be generated by many different methods, some not involving a camera, we can’t rely on cameras to generate unique identifiers, even though it is possible for cameras to do so. That is a shortsighted solution. However, the ability to feed unique identifiers into our cameras, with the cameras then assigning those identifiers to each image created, and with those IDs linked to remotely stored metadata—this is a viable solution.
Do you think there is a legal solution to allowing access to orphan works that appropriately addresses photographers concerns regarding orphan works? Or should the onus lie on photographers to implement appropriate business practices?
(Jeff Sedlik) Photographers do need to take responsibility for attaching rights information to their works in the most robust manner possible. Photographers also need to take responsibility for clear communication of rights to their customers, and for making rights information readily accessible and understandable during and after the license period. While these may or may not be legal responsibilities, these are professional responsibilities, and in fact are business imperatives if photographers expect to survive in our evolving marketplace. There is no easy solution to orphan works. In my experience, most photographers would likely support legislation allowing some degree of non-profit cultural heritage use of their images if orphaned, but generally object to legislation allowing any commercial exploitation of their images if orphaned. An expansion of fair use/fair dealing to allow certain usages of orphan works for cultural heritage purposes might be a path worth exploring.
Do you consider copyright metadata as a significant property of digital assets that should be preserved?
(Michael Steidl) The copyright law of some countries, e.g. the US and Germany, explicitly disallow the removal of metadata from the files of digital creative work which are used to claim copyright and being the creator of the work.
(Jeff Sedlik) Yes. We just need to implement a better means of transporting, storing and resolving that metadata.
Has the PLUS Coalition considered collaborating with any reverse image search engines?
(Jeff Sedlik) Yes, the PLUS Registry allows for the recovery of image metadata using reverse image recognition search. This is operational now for internal testing and will soon be available for beta testing. I will note here again that image recognition search does not solve the challenges facing photographers. Even if the orphan works issue were to vaporize, the underlying challenges would remain, and pose a very serious threat to the sustainability of the profession. See more detail in answers above.
Didn’t the DMCA law make stripping out metadata illegal?
(Jeff Sedlik) The DMCA [US law] makes the circumvention of copyright protection measures illegal, and that would arguably include the stripping of rights metadata. But the DMCA also relies upon the intention of the party stripping the rights. If that stripping is done with the intention of circumventing the protection afforded by the metadata, and with the intention of infringing or enabling infringement, that act would likely be a violation of the DMCA.
Other than changing a picture’s pixels a la the Marksta app, how can we enforce the retention of rights for binary content?
(David Riecks) The article given above in the resources (http://damcoalition.com/driecks/story/six-ways-to-keep-your-digital-images-from-becoming-orphan-works) gives five other ways.
(Jeff Sedlik) As mentioned above, providing contact info for a copyright owner is a big step in the right direction but does not solve the challenge faced by photographers.
How do you handle images that are taken with a phone and instantly uploaded?
(David Riecks) For iOS devices you can use Marksta (http://www.marksta.com/). Do test your workflow, as some of the services to which you upload will then strip the metadata from the image. Emailing an image from the Marksta App works, but with others, your mileage may vary.
(Michael Steidl) At IPTC, we are aware that more and more professional cameras have built-in Wifi or Wifi as add-on, and all smartphones have Wifi as standard component. Therefore we try to get in touch with the makers of cameras and smartphones to encourage them to add metadata features to their operating system already.
(Jeff Sedlik) The ability to pre-load your IDs or rights metadata and have your phone or camera load that metadata into each image captured is key.
Is there any ongoing pressure on the social media networks to rectify this?
(David Riecks) Since many of the social media networks do not charge you for the service, there isn’t a great incentive for them to change. The best pressure would be if a good percentage of their users kept asking why their service isn’t preserving metadata by default.
(Michael Steidl) At IPTC, we think the pressure which is considered as relevant by the social media network companies comes from their users. If some photographers would go to court because the network has stripped out metadata, this could change their approach. As IPTC got no response from the social media networks, we guess their reaction was: “It’s IPTC’s job to promote the use of their standard – this is only a marketing campaign – forget about it”.
(Jeff Sedlik) Given the circumstances, “pressure” will only come with litigation. Lacking pressure, the best way to resolve the issue is to structure a proposal that would allow the social media sites to increase their profits. This is the use of an incentive rather than (or in addition to) pressure.
Even if I protect my work, I can’t afford to sue anyone for using it. Is there anyway to be protected without needing lots of money?
(David Riecks) If you are located in the US, or have “registered” the work with the US Copyright Office, you will receive additional protections, such as being able to ask for statutory damages and attorney fees if you win an infringement case. Detailed info can be found in the ASMP Copyright Tutorial.
(Jeff Sedlik) I would estimate that 99..9999% of infringements are resolved without resorting to litigation. You want to stay out of court by whatever means necessary. In the US, a timely copyright registration is a great way to avoid litigation and settle infringements quickly and easily. Of course, you should take steps to attach your rights information to your images as well.
If each country comes up with its own laws, how will this ever be managed?
(Jeff Sedlik) If you are referring to orphan works, the best solution is to create a global network of interconnected registries. Those photographers who elect to register their works will then be immune to any orphan works legislation requiring a diligent search. If your question is more general, there are international treaties (Berne, TRIPS, etc) related to copyright, and most signatories to those treaties make attempts to abide by the terms. Copyright is generally enforceable across international boundaries. Some countries do not emphasize enforcement of intellectual property. (I’m being nice with that statement).
Has Adobe taken a stand against the social media networks on this issue?
(David Riecks) I’m not aware of any public statement that Adobe has made about this issue.
(Michael Steidl) No, and why should they? Their software is able to deliver photos, e.g. from Lightroom directly into Flickr. That means Adobe has done its job as expected. It’s more the job of the standardization bodies, like IPTC, and of photographer trace associations to protest against networks which strip off.
(Jeff Sedlik) We’ve found that Adobe is often responsive to respectfully stated concerns. Adobe is interested in serving the interests of its customer base. Remember that more than 90% of Adobe’s customers are not photographers.
I thought there was a law that prevented sites from being able to alter what was posted to them. Wouldn’t that apply here?
(Michael Steidl) The copyright law of some countries disallows the removal of metadata which are relevant for rights. But, the law of which country applies to a social media network? There has been a lot of discussion about that related to privacy.
(Jeff Sedlik) If this is a question about the use of images and the alteration of images and metadata when uploading images to social media sites, those sites put forth terms and conditions and rely on the users to agree to those terms, abide by those terms, and hold the social media sites harmless when users fail to abide by the terms. If the terms allow for the use and alteration of submitted images, and if users proceed to upload images, then the users are to blame. The primary problem faced by photographers is the unauthorized uploading of photographers’ images to social media sites. Again, those doing the uploading are to blame. I’m not suggesting that the social media sites’ terms are justifiable. I am suggesting that if you don’t agree with the terms, don’t use the sites.
Adobe should make entering metadata something that can be done from within the main PS UI. It’s buried or in Bridge, and out of sight, out of mind.
(David Riecks) You can access it from within Photoshop. It’s called File Info (Under the File Menu).