Lumen Announces New Features of the Database

Lumen is an academic research project, located in the United States, that studies the removal of online content. Our goals are to conduct and facilitate research on requests to remove content and links thereto; educate the public about different kinds of removal requests; and provide as much transparency as possible about such requests for interested parties.

Platforms, search engines, and others that receive court orders regarding links and content, Digital Millennium Copyright Act notices, and other forms of takedown requests voluntarily share those requests for inclusion in the database. We receive between six and seven thousand notices a day from the companies that voluntarily share notices with us, and as of May 1, 2019, the Lumen database includes over ten and a half million notices, regarding more than 3.3 billion URLs.

Lumen intends to remain a vibrant and valuable feature of the landscape with respect to research, journalism, and public awareness around takedown requests. We believe that the database has been a valuable tool for those involved in research and policy advocacy over the years. We also believe that it is both possible and necessary for Lumen to grow and improve.

One obvious way in which to do that is to expand the number and type of notices that the database receives and the range of institutions from which we receive them. Relatedly, as we expand further from the project’s origins as one focused on DMCA notices, and as we continue to receive more copies of a wider variety of takedown requests, we wish to remain conscious of balancing the interests of the research community with the interests of those sending notices to online platforms.

To that end, Lumen is announcing today some small but significant changes that will affect the amount of information within a notice that will be visible to a user of the Lumen site, unless that user is logged in. Specifically:

First, the URLs that are the subject of a takedown request will no longer be readable in their entirety. Instead, only the domains and top-level domain (“TLD”) of each URL will be visible. If there is more than one URL from a given domain and TLD, the notice page will indicate how many URLs there are.

So, for example, where now a Lumen user would previously see:

http://www.example.com/bad/url_11

the user will now see:

www.example.com — 1 URL

And, if twelve URLs from the same domain are the subject of a particular notice, the user will now see:

www.example.com — 12 URLs

Any file attachments associated with a notice (e.g., a copy of a court order or other document) will also be unavailable prior to a user’s logging in.

Second, below the truncated URLs in a notice, a user will also see:

Click here to request access and see full URLs.

Clicking this will take a user to a page where the user can enter an email and receive a link to the unredacted “full” version of the notice, including any file attachments, that will be good for only a limited period of time.

Third, researchers, journalists and policymakers may contact Lumen and request login credentials to the site which, once used, will allow them to see full URLs and file attachments for all notices and generate “permanent” URLs for the full versions of notices that may be used as hyperlinks in papers, articles and other materials requiring citations.

With these changes, interested visitors to the Lumen site will still be able to review URLs associated with a particular notice. Database users motivated further to do academic or policy-based research can apply for credentials with which to do so.

The Lumen team believes that these changes are an important step in the project’s ongoing efforts to achieve the balance we constantly seek to strike between our dual missions of transparency and research (on the one hand) and the strongly-felt concerns of those who send takedown notices (on the other). Lumen has always taken steps to avoid stigmatizing individuals involved with complaints, and the project has sought to avoid rendering judgments on the quality or provenance of the individual removal requests in the database. We have made (and continue to make) good faith efforts to redact personally identifying information from notices, where that information is not critical to understanding the complaint and its provenance. We have also taken steps to protect the legitimate privacy interests of notice-senders, some of whom would prefer not to be associated with takedown requests.

Now, as ever, we seek to best allocate scarce project resources to further the long-term goals of our effort — i.e., research and transparency regarding the takedown landscape.

The Lumen team believes that that the small but significant change described below will go a long way towards assuaging concerns and will allow us add more notice submitters to our list, increasing Lumen’s utility for everyone.