Relativity and AI - aiR for Review

Relativity's latest innovation is aiR for Review, which uses artificial intelligence to not only help find responsive documents, but also to describe why the document is relevant and cite to the parts of the document that make it relevant.

aiR is accessed in the Relativity mass operations menu.

. . . the litigation team enters a case summary using the same type of language you might use in a memorandum to describe the key players and key terms for a matter. A simple 2 to 5 sentence overview of the case should be provided:

The criteria to define exactly what should be regarded as relevant can be given as a description that you would send to another member of your team, rather than a series of keywords or sample documents.

When an individual document is viewed, the user can check a short summary of why it was found to be relevant, associated with the parts of the document that support this finding:

The process still requires a subject matter expert to not only devise the criteria used by aiR, but also to verify the initial results. A team of reviewers has to validate the findings generated by aiR on a statistical sample of the full document set.

When using aiR, Relativity recommends providing information clearly, but also as briefly as possible. The active voice should be used (so, "Mr. Johnson embezzled the bank's funds.", and not, "The bank's funds were embezzled by Mr. Johnson.") and double negatives should not be used. (Don't say, "Johnson can't hardly wait until the deal becomes final."). It is also necessary to avoid the qualifier, "including but not limited to".

detect whether or not text was generated by AI

It's becoming more and more common to come across text on the internet which appears as though it may have been written by AI. You may question the usefulness of the information, which may simply serve as clickbait, or want to avoid using it as a source in a legal filing. There are several services which can help determine if text was created by an artificial intelligence program. GPTZero, https://gptzero.me/ , will let you test a limited amount of text for free. It appears to be reliable.

Back on February 11, 2023, I posted about using ChatGPT to help solve litigation support problems. See this post.

I tested the answer that ChatGPT gave when I asked it to explain the EDRM.

GPTZero found that 11 of 18 sentences used in the answer were generated by AI, although oddly it said that there was an only 1% probability of the answer being AI text.

When I tested some text on a travel guide site that I thought was suspicious, it gave the same probability result, but more to the point did not think that any of the sentences were AI text.

Getting litigation support tips from ChatGPT

The open source artificial intelligence chatbot ChatGPT has been in the news quite a bit since it was released late last year. Whether you work with computers for your job or not, it's really something that you need to check out. If you're familiar with the range of assistance that AI devices like Alexa or Google Assistant can provide, you'll be in for quite a surprise at how much better ChatGPT is. It's really a game changer, and is pretty close to a substitute for attempting to find answers to problems by running Google searches.

You can create an account for ChatGPT and begin using it very quickly. See this link: https://chat.openai.com/auth/login

It does require that you provide a phone number, and comes with the disclaimer, "the system may occasionally generate incorrect or misleading information and produce offensive or biased content. It is not intended to give advice." It has limited knowledge of the world and events after 2021. The information that you provide it with will not be protected from disclosure.

I decided to begin testing ChatGPT usefulness to litigation support professionals by asking this question: "Can you explain the electronic discovery reference model to me?"

That's not a bad response, but including identification, collection, and preservation as part of information management is an odd choice, and I'd really wonder about how experienced someone was in electronic discovery if they chose to answer this way. It also would have been more impressive if ChatGPT provided an image of the EDRM.

I next decided to ask ChatGPT a more specific technical question. I wasn't really expecting it to provide a good response to a request for even a basic Regex script, but it surprised me.

There's no doubt about it, ChatGPT's answer, \b\d{9}\b , does work.

My next question asked the AI bot to craft a more complicated regular expression search:

It even gives the option to copy the Python code. I engaged with ChatGPT further asking for guidance on how to actually run the Python script, and it didn't let me down.

This is correct, although figuring out how to get Windows to recognize Python in Command Prompt after I installed it was a whole other problem that I decided to work out on my own. The solution should be part of an upcoming tip of the night. After I got Python to be recognized in Command Prompt I was able to successfully test out ChatGPT's script.

I don't like to promote the replacement of litigation support professionals, or this blog for that matter, but there's just no ignoring what a useful tool this is. I think we're at an inflection point. It will not be possible to work without AI tools like ChatGPT going forward, any more than it was possible to ignore the need for internet service in the late 90s, or the need for a smartphone after the introduction of the iPhone by Apple in 2007.

ChatGPT generated answers to my questions in a few seconds, and archived its responses. I can't imagine anyone in our field doesn't regularly expand upon their knowledge of Excel, document review platforms, or electronic discovery in general by running Google searches. Now, you'll be doing this with an AI tool that will provide more specific answers to your questions than Google or the whole world wide web ever could.