One of the biggest challenges I have faced in the past five or so years has been getting my clients and colleagues to share my enthusiams for AI or machine learning. 

I have found that "AI" can be perceived as something that is meant to take the human factor out of the equation, a tool that cannot be trusted, and something that might work as a back up if run in the background on a case. The reality of it is that all of these can be true, but so can the major benefits that I see for it; increase efficiency and speed of review, identify documents that may not have been considered in a traditional review and apply learning to a broader dataset in a way that is simply not possible without the technology.

Using a tool like TAR or CAL efficiently is largely down to the eDiscovery service provider's ability to deploy it, and understanding of the purpose of implementing it. If it is sold as a "magic bullet" it is likely going to leave a negative aftertaste with the users, but if it is discussed and intended as a tool to aid the review and bring additional support to the case team there is an increased chance that it is seen favourably. 

Another area that is sometimes misunderstood is the validation processes that are implemented at the end of using a tool like CAL. It is this process that proves how effective the solution has been in any one case. We must all accept that it isn't perfect and can only ever be as good as the information it is given but there are tolerences that we widely accept as showing a successful application of the tool - understanding this and being able to help our clients understand the sometimes "maths heavy" explainations behind it is a skill in itself. 

I believe that the best way to use AI and/or machine learning within the confines of litigation or an investigation, is to build a relationship with an eDiscovery provider who is able to guide you and understand the best uses and outcomes of the tools. It isn't a tool that will replace the human element of a case and shouldn't be treated as such, it is best used to supplement the human input and add efficiency. Someone who understands this and can help you and your colleagues make the most of it is going to be invaluable moving forwards as this becomes more and more common in the matters you work on.

If you would like to talk about how the use of AI can help, my (virtual) door is always open and I am always happy to discuss how something as powerful as AI can help, rather than replace, users.