One way to solve the federal records puzzle
If everything is a record, transparency and sharing become the rule and not the exception
Agency records officers must regularly decide what qualifies as a federal record. Maybe it would be better if they didn’t have to make that decision.
The U.S. Code formerly defines federal records in Title 44, Chapter 33, Section 3301 as all materials in “connection with the transaction of public business” (abridged). The collection and preservation of federal records originally focused on transferring physical paper records to the National Archives. The preservation of paper records is feasible, manageable and prudent because of the natural physical constraints of the medium. Unfortunately, electronic records, beginning most notably with e-mail messages, have exponentially increased the difficulty of preserving all materials, physical or otherwise, that relate to public business.
In addition, current laws say not everything is a federal record worth preserving, such as working drafts, transient data and coordination notes. This puts the burden of deciding what is and what is not a federal record up to each federal organization and thus subject to whim, opinion, Freedom of Information Act requests and, at worst, corruption.
In that light, let me propose a hypothesis: The notion of trying to distinguish what constitutes a federal record could be a dying concept — as opposed to just marking, compressing and saving all non-duplicative data.
That might sound like a radical change, but why are we afraid to say what is staring us directly in the face? Okay, I’ll say it: In public service, everything is a record.
I’m not saying that you need to expand manual processes tenfold to handle more volume. Instead, I am saying those manual processes for recording electronic information are not working, so you need to focus on automated processes across the entire data value chain from inception. That will work because automated processes can expand to handle much larger tasks.
Now let’s examine two of the valid areas of concerns about this hypothesis: volume and sunlight.
The argument about volume is simply that automation will drown our current processes and overwhelm our ability to save and search data. A counterpoint to this argument is how services such as Google and Microsoft Live essentially provide unlimited storage without sacrificing the ability to find the data at a later time. Though not perfect, those online services are harbingers of things to come — and I have not even brought in the big guns, metaphorically — of Moore’s Law and 64-bit computing. In addition, text data compresses well, and Web applications enable centralized data by default. Thus, I see the volume argument as solvable.
The argument about sunlight is that people will change their behavior if everything is automatically recorded for posterity. The counterpoint lays in analogies to other sunshine technologies, such as cockpit recorders and video cameras in police patrol cars. I don’t see pilots protesting their invasion of privacy from flight recorders, because the benefits outweigh the cost. And although police officers have less privacy on the job, those cameras have saved numerous officers from false accusations. Thus, I have faith that public servants will accept this as part of their duty and reap the benefits of automated record keeping.
Briefly, some of those benefits would include improved e-discovery, automated FOIA requests, true enterprise search, knowledge retention, improved public access to data, improved information sharing and robust, rules-based event triggers.
If everything is a record, transparency and sharing become the rule and not the exception. Also, as with video cameras in squad cars, it is better for those who have the public trust not to have to decide what is and is not a record. That is not a decision a public servant should have to make or want to make. In my opinion, this truly is a case where the “man-in-the-loop” is the wrong approach. Saving federal records is better left to the autopilot.