Integreon’s business is document review, for litigation, regulation and other purposes, including litigation content management, and compliance due diligence. Its business involves using a mixture of technology, human skills and well-honed processes to deliver document review services as quickly and cost effectively as possible. Its business model depends on delivering the agreed output on time and to budget. As a company, it therefore has a close interest in any developments which speed up delivery while maintaining accuracy.
Vince Neicho is VP – Legal Services at Integreon in London, after a long career as litigation support manager at Allen & Overy. He has written an article for Legaltech News [registration required] called Intelligently reinventing AI: using human intelligence to leverage the artificial kind (also available here on Integreon’s site)
The term “artificial intelligence” is widely used at the moment. There are a few software providers who can legitimately describe their products as bringing “artificial intelligence” to data and business problems, but the term is widely used to cover almost anything involving a computer. For example, the present issues in UK criminal law about disclosure of mobile phone data are often said to be soluble by “AI”, leading me to post this on Twitter recently:
Vince Neicho opens his article thus:
“Artificial intelligence” has overtaken “innovation” in just about every conversation, pitch or marketing piece. At the same time, there are still those who have absolutely no idea what it actually is. Clever computers? Trained computers? Active Learning? Statistical analysis? Complex algorithms? Smart devices? Robots?
The answer is, elements of all of the above—but it depends, so there’s a requirement to be specific instead of just spouting jargon.
Consistent with this opening, he stresses the need for human interaction with technology to “facilitate smarter working” and to “provide opportunities to delegate or outsource higher value work”.
This article gives some examples, and not all from the disputes context. Talking of the (often vast) task of extracting useful data from contracts he says this:
Automated metadata extraction and categorization technology can be used to speed up this process and reduce the cost associated with the traditional manual contract review. This technology auto-triages contracts based on certain business rules (e.g., the exclusion of expired, duplicate, unsigned or immaterial contracts), groups like-contracts together and prioritizes the most critical contracts for review before automatically extracting the metadata. Having set up and implemented the technology, humans can then efficiently work with the output to conduct their review, add their valuable work product, check for errors, such as false positives, and make key decisions in a structured manner.
The emphasis, therefore, is not just on speeding up the process, but on providing an efficient means of supplementing auto-extracted data with human input, checking for errors and making decisions.
One of the drawbacks of the hype to which Vince Neicho refers is that it somehow implies that the push of a button will solve the data problems, whether that be for litigation discovery, contract analysis, or the examination of mobile phone data in rape cases.
What is missing in this simple-sounding summary is the people and the processes which make the technology work. Integreon’s speciality is just this, and Vince Neicho’s article, with its focus on actual intelligence, makes it clear how that combination of skilled people, well-developed processes, and the right technology can get the job done.