Preventive risk system, Intraspexion, has increased the capabilities of its ‘pre-crime’ legal review tech.
Named by Artificial Lawyer – part in jest, and part seriously – as having the legal AI equivalent of the abilities seen in the movie ‘Minority Report’ based on the story by Philip K. Dick, the company founded by US-based Nick Brestoff, has sought to allow inhouse legal teams to scour internal documents and written communications using NLP to spot potential legal issues that could trigger litigation. In short: it’s a type of ‘pre-crime’ safety system, built into the company’s data and internal comms channels.
Interestingly, it’s actually one of the use cases corporates mention they do want to see addressed, often saying things like: ‘I’d like a system that helps me prevent legal issues, not just work more efficiently once a matter lands on my desk.’
‘The amount of money saved by giving an enterprise the means to flag a potential employment discrimation matter before it turns into an actual lawsuit is staggering,’ added Brestoff, Founder and CEO of Intraspexion.
For example, after a job interview two executives start an internal correspondence about why the deliberately rejected a certain applicant, mentioning issues that clearly would break employment law.
In this example there is no law suit yet, but if there was one such messages could be dragged up in discovery. The system at least alerts the company to what went on and perhaps gives them some time to address the matter, or prepare for what may be coming down the line – hopefully shortening the length and cost of any dispute, or preventing one from beginning entirely.
And in terms of how the tech works, F. Scott Barker, Intraspexion’s Chief Developer, said: ‘Intraspexion embeds the dtSearch Engine for its document filters and text retrieval capabilities. We have updated the software to include Office 365 connectivity with a console for either an automated (every night) or on-demand run. We use the dtSearch Engine to extract the enterprise email content. Then the system uses dtSearch to index and search the extracted content and passes the results to Intraspexion’s GPU instance residing in a cloud environment.’
Intraspexion’s technology makes use of a Deep Learning Engine to figure out which correspondence may indicate a future litigation risk. The Deep Learning Engine is a form of Artificial Intelligence consisting of a machine learning multi-layer neural network that Intraspexion trains to spot a specific category of potential litigation.