I was puzzled this week at the reaction to a bomb of a story by the Wall Street Journal. The paper’s rightfully cautious lawyers allowed it to go to press and declare that 131 federal judges had broken the law by hearing cases in which they or their families had a direct financial interest.

Photo: Uwe Kils, Creative Commons

Even though a bunch of judges admitted they had acted incorrectly, other than a couple of reports by The Hill and Esquire.com, the reaction was zero.[1]

The story and its aftermath yielded a couple of interesting lessons.

  1. If a story can’t be politicized these days, it goes away fast. The judges named in the story came from across the political spectrum. Obama appointees, Bush (both of them) appointees, Clinton appointees, Trump appointees. They should have recused themselves but didn’t. No political points to be scored here.
  1. As an investigator, the story was a wonderful illustration of the idea that the world is packed with information that nobody either looks at or analyzes. This story scratches the surface of what lies ahead as artificial intelligence (AI) both increases the amount of data we can search, and the speed with which we can analyze it. Imagine being able to see not just the part of the iceberg above the water, but the whole thing – right away.

JUST ASK FOR IT

In this case, the paper relied on a non-profit called The Free Law Project which put together the stock ownership data for the judges. How did the Free Law Project get this information, which judges are required by law to submit? They asked for it. That’s it. Now it’s a public database on their website.

There is a lot of other wonderful information that is there for the asking. Freedom of information requests at the federal and state government levels have given us all kinds of material over the years as we investigate.

Did someone really serve in the military? What did a charity say its mission was when it was established? What kind of things did a company import and declare to U.S. Customs? We’ve found it all by asking for it.

All sorts of other great information is there for the taking. We once linked two people using state-level lobbying records freely available on the government’s website. If your database doesn’t look at lobbying records, that should be the database’s problem, not yours.

PUTTING IT ALL TOGETHER

The brilliance of the Journal’s story wasn’t just the sea of data the Free Law Project gave them. It was the critical task of seeing what happens when you match one set of data with another. The second set, of course, was looking at cases all the judges had heard. When a party before the judge turned out to have been owned in part by the judge or the judge’s family (personally or in trust and known to the judge), that was where the law was broken.

One of the keys to a good investigation is how databases and other forms of information work together. I’ve written for years that we employ an “all windows open” approach. It’s no good running a search on Bloomberg and then checking it off your list as a source already consulted. If later in the investigation you find out in an interview or court record that your subject has a company you didn’t know about when the investigation started, you have to look at that company on Bloomberg again. [2] Think of it as having each information source in a separate open window on your computer. Don’t close all until the investigation is done.

It probably took the Journal a long time to do its analysis, but note that the story only handles cases between 2010 and 2018. Nearly three years of cases since then have been heard – thousands – and the “investigation” is far from complete.

THE FUTURE OF INVESTIGATION WITH ARTIFICIAL INTELLIGENCE

On a very basic level, the Journal used AI to parse all the data. Once parties and stockholdings are fed into a database, a computer can spit out conflicts in a second. Human beings need to carefully review the findings, but doing the entire job by hand would have taken years.

AI helps in processing time, but it also is helping to create searchable data sets that were never searchable before. I discussed this at length in a law review article a few years ago, Legal Jobs in the Age of Artificial Intelligence. Imagine being able to search in seconds the transcripts of every podcast or YouTube video on the internet.

That time is coming. More data, more analysis, more human hours to make sense of it. This Wall Street Journal story is just a taste of the future.

 

 

 

 

 

[1] The Administrative Office of the U.S. Courts, which runs the federal judiciary, told journalists it was looking into the matter. Many cases may have to be retried. It’s a mess for the judiciary which will play out over weeks, months, years, probably.

[2] My book, The Art of Fact Investigation discusses this concept, and I’ve included it for years in my courses and lectures to lawyers around the country.