3 ms·

Judges in Jeopardy: Could IBM's Watson Beat Courts at Their Own Game?

26 points by grellas 15y ago

sp332 15y ago
This is not very insightful, mainly because the author does not understand how the algorithms are chosen. The methods of attack Watson used in Jeopardy were chosen and tuned by the programmers. Even the automated training they gave it was programmed to find answers with an outcome the programmers chose. For example, when the author says Inasmuch as he makes errors, these errors are randomly distributed. His mistakes are not skewed due to political preferences, personal relationships, or other sources of human prejudice. Watson by design avoids the ideological bias of judges—which textualists so deeply fear— because, of course, he does not have any ideology of his own. Of course the computer will be looking for answers to questions that humans ask it, using techniques that humans programmed. Even if the humans introduce bias unintentionally, or accidentally recreate the effects of "ideology", the most you could say is that the computer is not "responsible" for skewing its answers. Of course the answers will be skewed, whether we know how ahead of time or not. There is no reason to think that the errors will be evenly distributed. http://en.wikipedia.org/wiki/Hacker_koan#Uncarved_block http://en.wikipedia.org/wiki/Hacker_koan#Uncarved_block
shawndumas 15y ago
print version: http://yalelawjournal.org/the-yale-law-journal-pocket-part/legislation/judges-in-jeopardy!:-could-ibm%E2%80%99s-watson-beat-courts-at-their-own-game?/print/ http://yalelawjournal.org/the-yale-law-journal-pocket-part/l...
roel_v 15y ago
While I'm the first to admit that I'm no expert on common law, I don't know of many proponents of textualism (outside of constitutional law) since Montesquieu. Seems to me like a rather vapid article that only applies within the constraints of very narrow definitions of some of the terms used in order to get to write a sensationalist headline.
- sp332 15y ago
  I'm not sure if I agree, but he does link to this PDF as corroboration: http://columbialawreview.org/assets/pdfs/106/1/Molot-Web.pdf http://columbialawreview.org/assets/pdfs/106/1/Molot-Web.pdf

26 points by grellas 15y ago

sp332 15y ago

This is not very insightful, mainly because the author does not understand how the algorithms are chosen. The methods of attack Watson used in Jeopardy were chosen and tuned by the programmers. Even the automated training they gave it was programmed to find answers with an outcome the programmers chose. For example, when the author says Inasmuch as he makes errors, these errors are randomly distributed. His mistakes are not skewed due to political preferences, personal relationships, or other sources of human prejudice. Watson by design avoids the ideological bias of judges—which textualists so deeply fear— because, of course, he does not have any ideology of his own. Of course the computer will be looking for answers to questions that humans ask it, using techniques that humans programmed. Even if the humans introduce bias unintentionally, or accidentally recreate the effects of "ideology", the most you could say is that the computer is not "responsible" for skewing its answers. Of course the answers will be skewed, whether we know how ahead of time or not. There is no reason to think that the errors will be evenly distributed. http://en.wikipedia.org/wiki/Hacker_koan#Uncarved_block http://en.wikipedia.org/wiki/Hacker_koan#Uncarved_block

shawndumas 15y ago

print version: http://yalelawjournal.org/the-yale-law-journal-pocket-part/legislation/judges-in-jeopardy!:-could-ibm%E2%80%99s-watson-beat-courts-at-their-own-game?/print/ http://yalelawjournal.org/the-yale-law-journal-pocket-part/l...

roel_v 15y ago

While I'm the first to admit that I'm no expert on common law, I don't know of many proponents of textualism (outside of constitutional law) since Montesquieu. Seems to me like a rather vapid article that only applies within the constraints of very narrow definitions of some of the terms used in order to get to write a sensationalist headline.

sp332 15y ago
I'm not sure if I agree, but he does link to this PDF as corroboration: http://columbialawreview.org/assets/pdfs/106/1/Molot-Web.pdf http://columbialawreview.org/assets/pdfs/106/1/Molot-Web.pdf