Artificial Intelligence Previews Alternative Final Judgments Revisited
Back in September, I posted on the topic of Artificial Intelligence Previews Alternative Final Judgments. My post was totally inspired by an article I had read in the Sydney Morning Herald about a software project in Australia. You can read my original post with the text of the Herald article here.
One of the developers of the profiled software, Professor John Zeleznikow of the School of Information Systems at Victoria University, recently e-mailed me with some comments on my post. I thank Professor Zeleznikow for his readership, for sharing his insights and for granting me permission to post his comments below.
A very interesting article.
I think you have misunderstood some of our research. Reading a
newspaper article does not give the full details of our work.First of all, we do not claim everything can be reduced to rules. In
some areas we use machine learning to try to understand patterns of
judicial decision making. And we accept there should be no
inferencing re child welfare or refugee law.In Australia, child support is mandated – the court has no discretion.
We have learned how judges distribute property.
We do not argue that child custody should be awarded to a parent
simply because that parent wants it the most and allocates the most
points to it. In fact we explicitly say interest based negotiation is
more relevant to industrial relations than child welfare issues.In Australia, parents do not have automatic rights to see for or care
for their children (the only test is the paramount interests of the
child). Nevertheless, some parents still negotiate as if they own
their children.
I stand by my original post, which, like my comments in this post, are about family law dispute resolution software generally and not about any specific piece of software in particular . As to the Victoria University software project, I have no specific firsthand knowledge or experience with it.
Still, I don’t think Professor Zeleznikow and I are in disagreement.
Rather, I think my original post focused on why I think the glass is half empty with respect to family law dispute resolution software in general. And Professor Zeleznikow’s comments focus on why he thinks the glass is half full.
I think we both agree on what it is that is in the glass though.
I look forward to following the progress, evolution and adoption of Professor Zeleznikow’s intriguing software in the field.
