Metadata Mysteries

Photo-Attorney Carolyn E Wright has stirred up a metadata controversy again with Is Google Stripping Your Metadata? posted a couple of days ago. In it she links to a couple of posts from the Gunar blog, Google in the hot seat for stripping metadata in image search results (May 27, 2010) and What should Google do about media metadata? (June 3, 2010)

As I’m sure we are all aware, the vital part of metadata for photographers is the copyright information which shows our ownership of an image, as well as our contact information. Google for its image search feature produces thumbnails of images from web sites, and in making those fails to include such ownership information. As Gunar points out, industry guidance – such as the Metadata Manifesto from the Stock Artists Alliance – is that ownership metadata should never be removed, and the technical means to transfer it when creating derivative files are well-documented and relatively simple.

© 2010, Peter Marshall
Katherine Grainger at the NPG. Thumbnail saved from Google search – no metadata.
© 2010, Peter Marshall

© 2010, Peter Marshall
Full size image found in Google seach – is saved from original location and so has full metadata.
© 2010, Peter Marshall

Gunar suggests that not only should Google always respect and transfer such information when it is present, but that it should also add the URL of the web page on which the picture is displayed. It’s data that Google obviously has and is currently in the link text on each image, along with the image URL. This would as stated be a very useful service, particularly for those older images put on line before we realised the importance of metadata and the threat of orphan works legislation.

As the post suggests, removal of metadata is illegal in the USA under the “copyright management information” (CMI) provision of the Digital Millennium Copyright Act, and according to leading copyright lawyers also in the UK. Perhaps it is about time that one of the leading photographers’ organisations in the US gave notice of an action against Google, who do rather seem to be dragging their feet over incorporating a few straightforward lines of code into their thumbnail production.

The same DMCA provision has also been dragged – perhaps rather less convincingly – into the dispute between Shepard Fairey and Associated Press over his use of an image of Barack Obama (news in the  BJP a week ago was the photographer Mannie Garcia has dropped his claim against AP) which is due to come to court in March 2011.

According to a post by Julian Sanchez on ars technica in March, AP are alleging that Fairey violated the DMCA copyright removal provision in making a copy of the work from which to produce his artwork. It isn’t clear what they mean by this, but Sanchez points out that “CMI embedded in a digital image as metadata, after all, will necessarily be omitted from a printed copy of the work.

Google are perhaps not the worst offenders of major online services. In April 2010, Jonathan Bailey reported on Plagiarism Today Flickr and Facebook STILL Strip EXIF Data. Flickr apparently now keeps it on the original uploaded files, but there is none on the other sizes that it generates. Of course most EXIF data isn’t a great loss, and what is important is particularly CMI data, most of which is IPTC data, but that probably goes the same way as EXIF.

Plagiarism Today also has some stock letters for making use of the DMCA to get content you own removed from web sites. It’s easy to do and there is useful guidance in DMCA Takedown 101, although I followed the perhaps more straightforward advice from Photo-Attorney Carolyn E Wright on NatureScapes. Unless you have some acceptable form of authenticated digital signature you will need to airmail or fax your signed take-down notice to the offending service provider’s DMCA agent.

It’s perhaps a symptom of the need to get more people to understand the need for metadata that almost all the web links I found when researching this article were about how to remove it rather than preservation.  Usually this was simply to reduce image size, although it was good to find the following in Yahoo! web developer’s Stoyan Stefanov’s Image Optimization, Part 3: Four Steps to File Size Reduction

“Important note on stripping meta information: do it only for images that you own, because when jpegtan strips all the meta, it also strips any copyright information contained in the image file.”

Also on my long trawl I came across a reminder that there can be privacy issues when EXIF data is included. If you are an illegal marijuana grower it probably isn’t a great idea to take pictures of your crop and upload them – even through an anonymous proxy – complete with EXIF geotags!

Leave a Reply

You must be logged in to post a comment.