2. The Big Data “Problem”
Artificial Intelligence directly connected to big
data
especially for predictive analytics
Doesn’t mean that big data is useless for
lawyers
it’s just not as robust as we would like for it to
be (yet)
3. we have a scarcity of data in the legal
industry
Smaller population -> less cases
Less litigious society -> less lawsuits
Less trials -> less written decisions
See The Big Data Problem for AI in Law, Sept. 11, 2016:
http://www.slaw.ca/2016/09/11/the-big-data-problem-
for-ai-in-law/
4. Current Limitations in Legal
Data Analytics
a rush to use Big Data can result in
‘overlooking a number of important
quantitative issues
bias in data sample
measurement error
questions of statistical significance
5. Why is this a Problem?
Data visualization can be an extremely
helpful tool to understand and
comprehend large amounts of data
Concern of falsely visualizing patterns which don’t
actually exist
without proper consideration Big Data can
be reduced to ‘quite useless or worse’
Erroneous and misleading conclusions
False cognitive biases that accepts small sample
sizes as a representation of the whole
6. The Proper Way to Deal with
Big Data in Law
Solution: use a ‘confidence interval to gauge the
margin of error for any data sample and
subsequent value
If Big Legal Data is used quantitatively, cannot be
done without use of inferential statistics
Similar to legal argument without case law or rules of
precedent
lacks meaningful point of reference, authority
Even then, results should not be accepted without
further enquiry
Robert J. Parnell,When Big Legal Data Isn’t Big Enough – Limitations in Legal Data Analytics, Sept. 26,
2016: https://settlementanalytics.com/2016/09/when-big-legal-data-isnt-big-enough-limitations-in-legal-
data-analytics/
7. How Do we Level Up?
In civil litigation, where much of the research in AI is
being developed, the majority of the useful data is held
privately
Information exists in silos and is not shared
Governed by confidentiality agreements
Protected as proprietary
Not properly maintained or aggregated
Either the information is shared, or 3rd party vendors
will aggregate with anonymization
8. Until Then, What is Big Data
Good For?
Doing what we do now, just doing it better
More efficiently
More thoroughly
More effectively
Allows small and solo practices to research or prepare on
the same scale as larger firms
Predictive analytics are still elusive
Unlikely to feature in practice prominently any time soon