« Do Search Firms Bias Inflate Compensation Benchmarks? | Main | In-house productivity muted by meetings? »
Artificial intelligence software may nibble away at lawyers
An article in the Economist (March 12, 2005 at pg. 34) described several “intelligent legal services,” including ones now in use for resolving Australian property disputes between divorcing couples, for assessing applicants’ rights to legal aid, and for deciding criminal sentences. Notably, no US lawyers or academics were cited, but a half dozen were from England and Australia.
The software uses expert systems, which are rules based inference engines (one has 94 different variables), as well as “machine learning” software. The latter allows the software to draw more and more accurate conclusions based on past decisions. The software adjusts itself after seeing the variables in past decisions and the outcomes.
I have written previously about document assembly, which offers a version of these capabilities.
Law departments individually or jointly could invest in this kind of software where they face repeated analyses of a similar kind. Is it too farfetched to have software do the initial analysis on whether a company must respond to this QUADRO request or whether an ad passes muster with the regulatory agency?
Think of the software being available to clients. I can foresee such software running on a company’s intranet and giving guidance on possible Foreign Corrupt Practices Act issues or non-discriminatory firing, to name two possibilities. The software would not reach a conclusion, but it could educate the client, sharpen the client’s appreciation for relevant factors, and train the clients. In very clear circumstances, it might answer the client’s question.
Posted on March 27, 2005 at 02:52 PM in Thoughts/Observations | Permalink
TrackBack
TrackBack URL for this entry:
http://www.typepad.com/services/trackback/6a00d834519fb069e200d83544912269e2
Listed below are links to weblogs that reference Artificial intelligence software may nibble away at lawyers:
Comments
You can follow this conversation by subscribing to the comment feed for this post.

