WHAT WOULD you do if opposing counsel came into your office in the middle of the night, rifled through your files, finding private comments, strategies and drafts of documents about your case? They might not need to. If you send interrogatories, pleadings, settlement offers or any other documents by e-mail, you may be providing this information without knowing it.

On Oct. 20, 2000, the Wall Street Journal announced that files created in Microsoft Word contain much hidden information that can be accessed easily-information such as who created the document and where it has been, as well as private comments and prior drafts. This information is stored automatically in every Microsoft Word document and much of it cannot be "turned off." The name given to this hidden, highly resilient information: metadata. While metadata has existed in Word documents for years, the article sent law firms everywhere clamoring for a solution.

Broadly, metadata is data about data, or rather the data embedded in documents or programs that explains the "why, where and how" about data, such as how and when a document was created, when it was edited and by whom. It can even include the edits themselves.

Web pages and many popular programs such as WordPerfect, PowerPoint and Excel employ metadata. With regard to Microsoft Word, metadata is the electronic document data embedded in all Word documents. Much metadata is created automatically and since it is not essential to most of what the average Word user does, most people remain unaware of it. Additionally, there is no limit to the amount of information that can be stored as metadata. Moreover, some metadata is very well hidden and because it is created in numerous ways, there is no single way to eliminate all document metadata. This should be of at least some concern to the estimated 70-plus percent of attorneys whose primary word-processor is Microsoft Word.

So what is the information embedded in your documents? According to Microsoft, there are at least 13 kinds of metadata in addition to switches or settings a user can activate that can create "hidden" document information. Types of metadata that should be of concern to most lawyers include document history, previous document content and embedded objects. Document history includes who created and edited the document, when it was created, how many times it has been edited and how much total time has been spent editing the document. Previous document content includes comments or edits and prior drafts of the document. This feature, while not automatically activated in Word, is useful in collaborative environments and might be activated by one person using the document without the other knowing. Embedded objects (such as spreadsheet information inserted into a Word document from Excel) can include the entire workbook. Consider the potential fallout if you intended to provide a client with his or her billing information and instead included billing information about all your clients.

(For a complete list of all 13 kinds of metadata, as well as several switches that create hidden data, see part II of this article.)

While metadata should be a concern for all Word users, lawyers are particularly vulnerable to it. Law firms produce much privileged work product such as briefs, contracts, memoranda, settlement offers and more in collaborative environments. Microsoft Word allows for easy collaboration of documents with multiple users. However, some of the same features that enable elegant collaboration also allow subsequent readers to see the entire document history.

There are some obvious ways this could impact your firm. The actual harm might be negligible if, for instance, a client merely sees a prior version of the final document you prepared for him or innocuous casual comments made about the case or a particular document. However, client relations could suffer if the document reports you spent less time editing it than you claim on your billing statement or if the client sees that the document was prepared first for another client. If you turned a document over to opposing counsel or a current client with the "track changes" feature option activated and the reader "rejects changes," prior versions of the document could reveal confidential information, thus breaching confidentiality. Another possibility is that you unintentionally disclose the identity of the people who helped edit the document, as well as the timing and effort involved in editing a document.

As more attorneys become aware of the problem, the more careful you need to be. Just because metadata hasn't been an issue yet is no reason to think you are exempt. The Oct. 20, 2000 issue of the Wall Street Journal mentioned the risks of metadata. Now lawyers everywhere are becoming more aware of metadata both as a potential risk and a potential resource.

First, the bad news: There is no way to manually remove all metadata from your Word documents. However, the good news is that you can minimize the amount of metadata contained in your Word files and greatly reduce the risk of a metadata incident in your firm.

Of course you can decide to never send electronic copies of your documents. For years, lawyers sent documents by FedEx or through the mail and it got there in one piece. You can fax documents to other firms (bearing in mind that if you use your computer as a fax machine, some programs can "fax" the complete document file to the receiving computer fax machine-not just an image of the document pages).

Some firms have elected to convert files before sending them, so that nothing is e-mailed or put on disk in its native Word format. The idea is that converting a document to PDF or text format (.txt) will capture just the viewable words into the new file and thus all metadata will be removed. But PDF is not editable and text format does not retain formatting or style selections. Converting the document to HTML is not an option; unfortunately, Word has such a good HTML converter that the metadata will be stored in the HTML document as well.

For ease of use and compatibility, some firms opt for an automated process by creating their Word documents with macros that eliminate much of the unwanted metadata. If you do create a macro, you will want to consider factors such as: What information is still accessible about the new document? How easy is the macro to use? Will it alter existing documents? And will it require significant customization to use?

There is also an excellent freeware plug-in for Microsoft Word that eliminates much of the more visible metadata from your document. You can download this program at http://www.payneconsulting.com. The software is free, at least in part, because Payne Consulting Group does not claim that it eliminates all metadata. You can install this or any similar metadata utility. If you do, you still need to consider what information remains in the "cleaned" document, as well as whether the utility will run automatically or whether it will be necessary for the user to clean the document after each use. This can be an issue in shared environments where another user might access or leave an unclean document or simply might be unaware of the need to clean the document.

Less elegant, more extreme solutions include running the document through a text- or hex-editor to ensure there is no compromising metadata in it before you send the document and saving the information in another format, like a text file, then copying it into a new Word file to affect final formatting changes.

The bottom line is that whatever word-processor you use, you should become very familiar with its editing and revision settings. You should also raise these issues with other members of your firm and technology support personnel.

Home | About the Bar | For Members | For the Public | Legal Links | Publications | Online Store
Lawyer Referral Service | Law-Related Education | NHBA•CLE | NHBA Insurance Agency | NHMCLE
Search | Calendar

New Hampshire Bar Association
2 Pillsbury Street, Suite 300, Concord NH 03301
phone: (603) 224-6942 fax: (603) 224-2910
email: NHBAinfo@nhbar.org
© NH Bar Association Disclaimer