06 Nov The Future of eDiscovery: AI and Machine Translation
In 2012, the Rand Corporation found that an eye-watering 73% of e-discovery costs are attributed to document review alone1. And that’s without considering the manpower needed to even get to the document review stage. Identifying, collecting and processing data each have their own challenges, and the manual processes behind these tasks are equally daunting in any given case.
In the era of big data, eDiscovery has its golden moment to marry up with technologies that take away the sheer time and cost involved in electronic discovery for legal proceedings.
An early 2018 survey by Consilio2 found that 93% of lawyers surveyed thought that artificial intelligence (AI) will prove helpful to the legal sector.
With technology-assisted review (TAR) already being advocated by courts and gaining huge traction among law firms worldwide, lawyers see the clear benefit from this technology, witnessing its increasing use in the past 5 years.
Predictive coding and computer-assisted review have also taken off in recent years – but lawyers are hungry for more.
In comes machine translation
In the same way that TAR initially had a slow uptake, machine translation has – in recent years – been in the shadows. Still dogged with a reputation for laughable quality and shaky security, machine translation has understandably had trouble in getting off the ground in the legal world.
What most legal practices aren’t aware of is the sheer revolution that has taken place in the language industry in recent years. Even non-traditional, but well-known players, such as Amazon and Facebook, have got in on the act and pumped millions into research and new machine translation tools.
The results have propelled machine translation into the 21st century – a new algorithm known as “neural machine translation” – has dramatically improved results, and new providers are hosting totally secure solutions.
This is a well-kept secret among many law firms who are catching onto these developments and handling multilingual eDiscovery cases with more ease. Imagine having a case with several languages to sift through – it takes an already complicated process and adds a healthy dollop of complexity on top of that. But machine translation is quietly revolutionizing the top eDiscovery players. What was once a headache is quickly becoming another run-of-the-mill case thanks to advances in technology.
A question mark on security
Security has, is, and always will be a top concern for lawyers. Nowadays with the GDPR directives, protecting data is a number one priority. Computers are starting to ‘think’ more in the way that people can, not by simple rote learning but by neural network architectures – a kind of lateral thinking for machines. As natural language processing algorithms have improved, it’s possible to ask computers more complex questions – and to get the right answers. Big data is leading to even bigger advances, and so far, AI, TAR and machine translation – through a specialized language service provider – are keeping up with the needs of eDiscovery cases and the legal industry.
Staying ahead of the competition
One thing that’s clear is that technology has always given a competitive advantage to those firms that can afford it. With eDiscovery traditionally being a heavily manual procedure, secure technology is warmly welcome. Machine translation, while relatively new on the scene, is set to become a game changer for the industry looking to find an edge over other providers, as the uptake of TAR and predictive coding reaches its peak. But with security and data protection at the top of the agenda, don’t plug data into Google’s public engines – make sure you choose a secure provider to stay ahead of the competition.
1Source: The Rand Corporation monograph – “Where the Money Goes – Understanding Litigant Expenditures for Producing Electronic Discovery”.
2Source: Consilio press release, 27.2.2018