We asked OpenAIs’ ChatGPT about Filevines’ Plans for a New Document Standard and the Future of Interoperability in Legal Technology …
Does anyone remember what life was like Before Google (BG)? Finding stuff on the internet really sucked. The browser was the focus and people used a product called Netscape and went to Yahoo’s portal to find information. Your internet access was even mailed to you in a silver CD-ROM case from AOL.
ChatGPT is giving us a glimpse into the future of digital interactions with humans. ChatGPT is not based on classic search, where each search is run independently; rather, ChatGPT leverages a concept called “conversational search.” Conversational search leverages past search queries as context for future search queries, which ultimately leads to more relevant search results. Alex Kantrowitz, a CNBC technology contributor, even questions, “Why Google Missed ChatGPT?” It took 75 days for Instagram to get to 1 million users, ChatGPT reached the million mark in five days. This is our Google Launch Moment!
I know there has been a ton of stops and starts regarding AI for a very long time. I admit to buying Prolog back in the day and, in 2022, the “artificial intelligence” programming language turned 50. AI has been incorporated into everything digital. In fact our iPhone tactile keyboards actually change the size of individual keys based upon probability of a letter being typed. At a minimum, we have to admit to being very close to an actual “conversational AI.” Zach Abramowitz is an investor in and has written thoughtfully about Legal Technology and Investing for years. He has also posted about the OpenAI product in his Substack which is worth a read, “Is ChatGPT Going to Change Everything?”
OpenAI has already led a $5 million funding round for a stealth company named “Harvey.” Harvey was founded by Winston Weinberg, a former securities and antitrust litigator at law firm O’Melveny & Myers, and Gabriel Pereyra, previously a research scientist at DeepMind, Google Brain (another of Google’s AI groups) and Meta AI.
“Our product provides lawyers with a natural language interface for their existing legal workflows,” Pereyra told TechCrunch in an email interview. “Instead of manually editing legal documents or performing legal research, Harvey enables lawyers to describe the task they wish to accomplish in simple instructions and receive the generated result. To enable this, Harvey leverages large language models to both understand users’ intent and to generate the correct output.”
We decided to see how smart the ChatGPT really is, so we asked it a couple of questions about interoperability in legal tech and Filevine’s new document standard that is supposed to rival PDF files and Microsoft’s DOC file format. The .vine “next generation” standard was announced on November 1, 2022.
Imposing a standard has not worked in the nearly $4 Trillion US healthcare industry and it certainly won’t work in the global $900 Billion legal industry. Bob Ambrogi editor of LawSites and columnist at Above The Law, has characterized the Filevine effort to create a new standard for documents as “misguided.” We also have a bias that agrees with Bob Ambrogi’s early November assessment.
I listened to the November 15, 2022 “Scale Document Generation and Automation with Filevine’s Document Assembly Webinar” narrated by Filevine employees. It is clear that the .vine format will not displace PDF or DOC files anytime soon, but it is definitely a conversation starter for interoperability in legaltech.
During the presentation, Filevine identified a hypothetical bank client that commented on a legal document generated by Filevine in a .vine format. The screengrab of a shared document is depicted above. Filevine is clearly doing an excellent job managing versions and authorized access. In fact the list on the right side of the screen shows a list of users that are authorized to view, and potentially edit, the document in .vine format.
In another hypothetical not specifically addressed on the webinar, it is possible that doctors, clinicians, and case managers may need to have document access as well for a workers’ compensation, personal injury, or medical malpractice case. For instance, connecting with Filevine’s authorization to create an interoperability standard would allow the authorized physicians in Giupedi to access a medical record.
Creating a true standard would allow this interaction to happen, or at least provide a roadmap for organizations integrate their products without bouncing between products or entering duplicative data. What if the Commercial Bank Client in the hypothetical was already using Litify or Clio for it’s in-house counsel? In fact, Litifys’ President, Ari Treuhaft, has already discussed, “Too Many Legal-Tech Tools Don’t Jibe with One Another.” Clearly, some type of standardization makes sense through APIs or document formats, and Filevine is definitely familiar with APIs.
One of the questions during the Filevine webinar stood out in our minds. A client asked if the documents would come in a template library. At first glance, providing a document library would make sense, but having clients create the specific documents across the diverse legal spectrum makes even more sense. The way the documents are assembled could also be a great fit for a product, of future iteration that works with ChatGPT. OpenAI has already created APIs that allow ChatGPT to work with customized content, also known as a corpus in “AI lexicon.” It is possible to imagine using a vast library of legal documents as a custom legal corpus. Onboarding and ingesting all of those .vine documents from thousands of users, all of a sudden, becomes much easier for a machine learning bot. Having the vast majority of the documents de-identified, with personal information removed, doesn’t hurt either.
This is not the realm of fantasy. Articles have been published from nearly 9 months ago discussing Beta versions of GPT-3 which is a more functional version of the popular ChatGPT. On March 30, 2022 a Medium post by Luciano Briata discussed the creation of a custom chatbot using GPT-3. GPT-3 is considered ChatGPT’s big sister and has 175 billion parameters versus only 20 billion in ChatGPT. The content data from .vine, PDFs, DOCX, etc… will be integrated at one level or another into a legal corpus in the near future as demonstrated by the Harvey investment.
At Giupedi, we’ve been working with artificial intelligence and machine learning technology for workers’ compensation and personal injury legal documentation. We have been utilizing a documents that are utilized in medical indexes as the primary content for our work. Much of the Giupedi work to data has been mostly a backend development effort. What we are seeing with the ChatGPT and GPT-3 is what is possible at the frontend interface.
Conversational AI can change the literal “face” of legal interactions for attorneys and clients alike. Like, Filevine, we believe the content of the medical/legal data that we are processing could be very imporant for new products to be introduced in the near future.
It’s very possible that this is just a first shot across the bow to create a standard document format and get the conversation started. The future, according to the OpenAI will depend upon the input of multiple stakeholders and lots of content created by lawyers, to be monetized by artificial intelligence bots like ChatGPT.
Edward Bukstel
Gerard Foti, DO, Orthopedic Surgeon 25+ years in Workers’ Compensation
