Old Dog, New Tricks

The other day I was working with a Master Carpenter.  What impressed me was how many tools he had in his shop.  Some I could identify, others not so much.

As an omniscient law librarian, I have many tools at my disposal, too, and have used many legal databases in my time teaching and research law and legal things.

The problem with using legal databases like Westlaw or Lexis or Bloomberg Law or Fastcase, Anylaw.com, etc. is that you (as in the individual) have to come up with a search string(s) that you hope will dig up whatever it is you're looking for.

What is a search string?

While you can always run searches using natural language (like what most people do in Google), a search string typically uses more complex terminology.  It's like it's own programming language and when you get into it, it really helps to narrow down your search results to the point where you can really dial into what it is you actually need.

Formally called Boolean search syntax, a Boolean search syntax uses terms and connectors to help narrow search results.

Boolean search connectors common to most all browsers and legal databases include:

 

 

And if all you're going to use is your common, everyday browsers, these will work great.

 

However, if you're ever going to step up and play with the big dogs, you're going to need to know how to develop a proper Boolean search syntax using one (if not all) of the more powerful legal databases.

 

My personal favorite legal database is Westlaw.  Powerful, tons of features and, best of all, easy to teach and use.  A bit on the pricey side if you're looking to get personal access, but it is hands-down the best resource when it comes to searching all things legal.

 

Back in the day, I was pretty good at developing my own search strings.  When I got stumped, I'd waste not a second and call 1-800-Westlaw and speak with a Reference Attorneys (bunch of lawyers who ate, slept and lived Westlaw 24/7) to get help drafting a search query.

 

However, with the advent of Artificial Intelligence (aka AI), coming up with a search string is easier.  So, let's say I searching in Westlaw for caselaw for/to:

Create a search string using terms and connectors seeking cases since 2015 where an amusement or theme park was held liable for injuries to guests

In this case, I used Chatgpt and retrieved the following result:

("amusement park" OR "theme park" OR "water park") /p (guest OR patron OR visitor OR invitee OR customer OR "paying customer")
AND (liab! OR "held liable" OR "found liable" OR "liable for" OR "judgment for plaintiff" OR "verdict for plaintiff") /s (injur! OR accident OR mishap OR death OR "wrongful death" OR harm OR "personal injury" OR "bodily injury" OR "catastrophic injury" OR "traumatic injury" OR "fall injury" OR "ride injury" OR "slip and fall")
AND (neglig! OR "duty of care" OR "breach of duty" OR "premises liability" OR "premises defect" OR "failure to warn" OR "inadequate security" OR "inadequate maintenance" OR "dangerous condition" OR "attractive nuisance") /p ("amusement park" OR "theme park" OR "water park") DA(aft 01-01-2005)

So, here we have a number of connectors like:

• OR (give me A or B; more expansive)
• AND(give me A and B (more restrictive)
• /p (A within the same paragraph as B - expansive)
• /s (A within the same sentence as B - bit more restrictive)
• ! (acts as an extender.  So, Act! retrieves Act, Acts, Acting, Acted, Acted, Actor, Actress...)
• ( ) (keeps related concepts together)
• "  "  (returns two or more words together in a certain sequence)
• DA (refers to a date filter)
  

With a date filter of 2015-2025, we'll see things we are looking for.  Or, at lease we can get to a starting point.

Lexis is another legal database.  Not my favorite as it is a pain to teach and it is still has some clumsy elements.

Using the same criteria as above (create a complex search string using terms and connectors seeking cases since 2015 where an amusement or theme park was held liable for injuries to guests) and using Chatgpt, our search string looks like:

("amusement park" OR "theme park" OR "water park") w/15 (guest OR patron OR visitor) AND (liability OR "held liable" OR "found liable" OR "judgment for plaintiff") w/15 (injury OR injuries OR accident OR death OR "wrongful death" OR harm OR "personal injury") AND (negligence OR "duty of care" OR "premises liability" OR "attractive nuisance") AND NOT ("judgment for defendant" OR "defense verdict" OR reversed)

Again we have a number of connectors like:

• AND (A and B - less restricted)
• OR (A or B - more restricted)
• AND NOT (A but don't include any results with B - even more restricted)
• w/# (I want A within a certain number of words)
• (  ) (keep this cluster of terms together
• "  " (I want these words in just this order)

Also to note is that Chatgpt added some legal terms like "premises liability," "attractive nuisance," and "wrongful death" suggesting that it is trying to give additional suggestions as users conduct a search.

Note also that this search query isn't as long or complex as the one for Westlaw.  That doesn't mean it's wrong or anything - it's just different and it's how Chatgpt interpreted what I was asking.  Change the query, and I'd get another result.

The bottom line here is that AI had really changed how law people work.  No longer do we have to hunt and peck around hoping to hit pay dirt.  Now we have AI search engines which help to cut down the time that is used to take to get started.