Alex G Smith asks if we should stop using the AI acronym and use something else in LegalAI
So Artificial Intelligence sounds awesome right ... it's intelligence that is artificial. It nicely evokes images from our favourite sci-fi films - especially when we blank out the dystopian plot (made by humans usually messing up technology with their own stupidity) and focus on the cute robot in the film. So great we now apply AI to every industry and we have a bold future full of cool new things, laughing at the way we do things now and have futurists dumbing down complex technical and mathematical approaches. But we're excited ... add in some start up fever, some over-optimism and loads of VC cash and we have ourselves a movement complete with pictures of robots everywhere. AI - it's even a verb now, as in, ‘we couldn't be bothered with accurate analytics for the last 10 years but hey we can AI the figures and it will fix our problem overnight’.
But what if it was called something else, what it really is right now in ‘LegalTech’ - would it attract so much fuss? Let's give this a go.
Introducing the acronym Legal D.C. or #legaldc
So most of the ‘AI’ tech in Legal at present is really data clean up and data structuring tools. It's doing its best to add a semblance of structure to hugely unstructured data by tagging, extracting, categorising and finding ‘things’. When legal digitised (firms, governments, courts, businesses, agencies, regulators) they digitised into Microsoft Word. Into forms, tables, essayed documents for judgments, word styled contracts, paragraphed regulations. Very few exceptions exist and they were due to people with vision like John Sheridan at legislation.gov.uk or the team at Propylon working regulations and legislation into xml. So basically before anyone does anything clever we need to do data clean up ... from experience in publishing this will be an iterative process of waves of clean up, structuring and rework and enrichment that may take a decade. It will be a machine-assisted but human-assured process. Given the accuracy of many of the tools it will be circa 40% tools, 60% humans. Or alternatively we could do what publishing did and ignore the tech, just outsource it. We'll need to talk xml, DTDs, schemas and write clean up instructions and send domain experts to outsourcing centres to train teams of data converters.
So let's get the Legal Data Clean-up movement started... D.C. as the acronym for the press and hype headlines ... oh, they lost interest and have moved onto a shinier shiny bauble and quick fix tech. Here's to legal data clean-up. The hard yards of the next decade digitising law.
Sadly, I seem not to be invited to speak at any conferences and am left sweating the hard #unsexy yards. Ah well, it's what a lot of industry have had to do ... no magic pills I'm afraid.
Alex G Smith is Innovation Manager at Reed Smith LLP (Senior Product Manager, Service Designer, Legal Data Architect). Alex has spent 20 years developing digital tools for lawyers and was running big data, search and ‘AI’ projects before it was cool, at LexisNexis and now at a global law firm. Alex has worked with some amazing and talented technical developers and loves the opportunity which technology and data can bring to legal.