If Rosanna Garcia had not rented out her property on Airbnb and if the strangers who stayed there had not damaged it, then this professor of marketing may never have created Vijilent. This is because the experience drove her to explore ways to find informative data on people, even if they were strangers, so that she and others would not have the same sort of surprise.
Vijilent is a natural language processing (NLP)-powered system that trawls through social media and other public data to build a ‘Data Portrait’ of an individual. This can then be applied in legal situations, such as jury selection in the US, but also could provide help in analysing other participants in court hearings.
The North Carolina company’s genesis goes back to 2015 when it had a more general background checking focus, but when Garcia was asked by a jury consultant about if her system could be of help in terms of jury selection the start-up began to focus more on the legal space, which is where it is now today.
Along with its own technology that the company has developed, Vijilent draws on IBM Watson’s sentiment analysis models, called Big Five.
Big Five is described by IBM as: ‘a model to generally describe how a person engages with the world. The model includes five primary categories, or dimensions:
‘We see this of being of use to litigation lawyers, private investigators and the justice sector. It could be of use in cases related to family law,’ Garcia adds.
She also notes that as more examples are tested (and you can request a profile of yourself here) then Vijilent will be able to improve.
The move into the legal space by the company appears to mark a growing interest in using analysis of public information, such as social media e.g. Twitter, LinkedIn, Facebook, to conduct personality profiling for legal matters. Last month Artificial Lawyer wrote about another new company working in the same field, Voltaire, though which takes a different approach in some aspects of the analysis and data presentation.
This new trend is all very exciting, but it has to be said, it does raise some concerns. Garcia notes that there are limits to where such personal data analysis can take place and where it can be applied. For example, crawling over personal data to build personality profiles that are used then by private businesses may not be possible in the EU. While in the US, regulation there may also restrict such ‘profile trawling’ in areas such as housing and employment decisions.
Then there are issues in relation to the impact of making use of social media posts as a basis for judging someone’s character. The main issue it appears is how far such systems present a truly broad, balanced and detailed picture of a person, rather than being swayed by a number of ‘hot button’ key words that might occur in a small sample of a person’s social media activity.
I.e. the proof of which personality profiling systems are the best will be in whether they can produce a clearly accurate picture of a person, rather than just a snapshot of a rush of tweeted thoughts over a short period.
As one can see, this is likely to be an area of growing interest, given the massive amounts of personal data we all leave ‘lying around’ for people to examine. AI technology, such as NLP, can help to reveal many things in unstructured social media data, blog posts and other public communications, but the questions remain: how sophisticated is the NLP that people are using and how are the extracted pieces of information being interpreted?
We can be sure this is just the beginning of the debate and to some extent the birth of a new industry that is AI-driven personality profiling in the legal sector.
If you have views on this issue, or would like to tell Artificial Lawyer about your company’s ventures into this space, please say hello, it would be great to hear from you.