4/24/13

Fastcase's Bad Law Bot: "Big Data Applications For Legal Research"

The bad boys of legal research, Ed Walters and Phil Rosenthal of Fastcase, are once again looking at unique ways to look at legal information and create new methods to cull that information. In the latest iteration, they have come up with a way to use an algorithm to identify court cases with negative treatment. They are calling this enhancement, "Bad Law Bot", not to be confused with J.J. Abram's movie production studio called Bad Robot.

The idea of algorithmically setting up a way to identify 'bad law' has been floating around since the idea of placing legal decisions in database began. When I was at the Oklahoma Supreme Court's OSCN.NET, we dreamed of doing exactly this same type of identification of bad law, but simply did not have the technology, expertise, or guts to take on that challenge. Looks like Walters and Rosenthal are stepping up to the plate to take a swing at it.

Ed does list a couple of caveats, that should be expected when you use technology to replace humans on decision making processes like this:

  1. It's an algorithm… thus the "bot" name
  2. If you see that Bad Law Bot has presented negative treatment, then that means there's a good chance the case has probably been overturned, however if Bad Law Bot doesn't show negative treatment, that doesn't necessarily mean the case is 'good' law. You should double-check with Shepards or KeyCite.
Despite these caveats, the fact that Fastcase is willing to go out and present something like this to its users shows that they are ready to test the boundaries of what you can do with legal information, technology, Big Data concepts, and the guts to go out and actually do it.

Bad Law Bot is available starting on April 25th, and the press release from Fastcase is included below. Also, Ed Walter's introduces the product in this two-minute YouTube video.







Fastcase Enhances its Authority Check Citator Service

“Bad Law Bot” Uses Big Data to Identify Negative History for Judicial Opinions

Washington, DC (April 25, 2013) – Legal publisher Fastcase today released an algorithmic enhancement to identify overturned or reversed cases in its Authority Check system – Bad Law Bot. Bad Law Bot uses algorithms to identify court cases that are cited with negative treatment and to alert researchers of a case’s negative citation history.

The Bluebook manual for legal citation requires that, when courts cite a case that has been overturned or reversed, they say so right in the citation. Judicial opinions, and particularly their citations, are full of this kind of “big data” about which cases are still good law. Bad Law Bot scours all of the citations in judicial opinions. When the opinions cite a case as being overturned, Bad Law Bot flags the case for Fastcase users, identifying negative history as reported by the courts.

“Fastcase’s Authority Check feature is already a very powerful tool for identifying whether your case is still good law,” said Fastcase CEO Ed Walters. “Authority Check includes data visualization tools to see the later history of cases, citation analytics and filterable lists of later-citing cases. The addition of Bad Law Bot, to help identify negative history, is a major step forward. This is the first of many additions to Authority Check that we’ll roll out over the next year.”

The new Bad Law Bot feature helps users identify negative treatment of the cases judicial opinions. However, because it only reports what cases say in citations, researchers should rely on Bad Law Bot as an aid to identifying negative history, not as a comprehensive guide.

Since 1999, Fastcase has been building smarter research tools for understanding the law. In 2012, the company launched eBook Advance Sheets available for the major eReaders (iPad, Kindle, Android, and Nook). 

In 2010, Fastcase was the first company to launch an app for legal research, and later, the first company to launch an app for iPad. The American Association of Law Libraries named Fastcase for iPhone the 2010 New Product of the Year. In 2011, Rocket Matter named Fastcase’s apps for iPhone and iPad the Legal Productivity App of the Year and the company furthered its mobile market presence by debuting the Fastcase for Android app in 2012. Lawyers on the go appreciate Fastcase Mobile Sync, which allows full integration of its mobile apps with the desktop version of Fastcase.

Fastcase has gained very strong momentum in the legal research market and continues to challenge the norm in legal publishing and legal technology. Fastcase was voted #1 in Law Technology News’s inaugural Customer Satisfaction Survey, finishing first in 7 out of 10 categories over traditional research providers Westlaw and LexisNexis. Fastcase has introduced new opinion summaries, Fastcase Cloud Printing, and has been named to the prestigious EContent 100 list of leading digital publishing and media companies alongside Google, Amazon, Apple and Facebook for two years in a row.

For more information on the Bad Law Bot feature, visit the Fastcase Legal Research Blog at www.fastcase.com/blog and watch this video: http://youtu.be/ZsKu7FoO2Ns.

About Fastcase

As the smarter alternative for legal research, Fastcase democratizes the law, making it more accessible to more people. Using patented software that combines the best of legal research with the best of Web search, Fastcase helps busy users sift through the clutter, ranking the best cases first and enabling the re-sorting of results to find answers fast. Founded in 1999, Fastcase has more than 500,000 subscribers from around the world. Fastcase is an American company based in Washington, D.C. For more information, follow Fastcase on Twitter at @Fastcase, or visit www.fastcase.com.


###

Bookmark and Share

2 comments:

Michael Ginsborg said...

It would be useful to see how closely the negative treatment results compare for Bad Law Bot, KeyCite, and Shepard's. Which one has the highest recall and precision? The results may well show that each of the three services should be used to supplement the other two.

In 1998, Elizabeth McKenzie reported that KeyCite and Shepard's editors assigned treatment notes to citing cases. (http://lsr.nellco.org/cgi/viewcontent.cgi?article=1058&context=suffolk_fp) A review of the relevant literature may show that KeyCite and Shepard's now use a significant degree of automation to assign treatment notations. So a "negative treatment" comparision of the three services would seem likely to involve a test of the effectiveness of the algorithms used. Moreover, Susan Nevelow Mart has already produced evidence on precision rates in automated headnote assignment for KeyCite and Shepard's. (http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2188541) A comparison of the kind proposed here therefore has precedent and may have promise.

Darren Wells said...

I agree Michael, in my opinion- the more tools the better. Whenever something is automated, there is always chance for error. As long as it is used in conjunction with other methods, it's seems like a great opportunity. Definitely has potential- I can't wait to see how that technology would be used in a wide scale situation.

 

© 2014, All Rights Reserved.