Thanks for bringing that to my attention. My initial thought is I hope the S.Ct. bans use of secret algorithms. Don't see how it squares with the 6th Amendment. Perhaps someone can provide summary legal analysis. ________ It doesn't really matter what Phil is saying, the music of his voice is the appropriate soundtrack for a bicycle race. HTupolev
Thanks for bringing that to my attention. My initial thought is I hope the S.Ct. bans use of secret algorithms. Don't see how it squares with the 6th Amendment. Perhaps someone can provide summary eval analysis.
Yeah I got no problem with using an algorithm, but keeping it secret isn't copacetic.
One of the problems with allowing the use of this sort of evidence is that the expertise to question the algorithm is probably hard to find as well. It was like this in the early days of DNA testing. An "expert" would testify that "there is only a one in million chance" that the DNA would match in the general population. But if they ran it through a database of 5 million records you could get 5 matches etc. If the algorithm remains propriety until it is challenged the only "experts" on its use likely will be employees of the company or the developers. This leaves the defendant naked to defend against it as finding a expert to challenge the finding is almost impossible. This should not be allowed. They constantly try to escape from the darkness outside and within Dreaming of systems so perfect that no one will need to be good T.S. Eliot
I would think that would have to be demonstrated before using it in the first place.
Nah. It's a mathematical model. All you need to do is get 97% of scientists to agree to it.
Again, only half joking.
What if the algorithms are demonstrated to work? What if they work as good or better than human judgement? What if they predict recidivism more reliably than human judges do? Shouldn't policy be driven by science, and all that?
"People think it must be fun to be a super genius, but they don't realize how hard it is to put up with all the idiots in the world."
I would think that would have to be demonstrated before using it in the first place.
Nah. It's a mathematical model. All you need to do is get 97% of scientists to agree to it.
Again, only half joking.
What if the algorithms are demonstrated to work? What if they work as good or better than human judgement? What if they predict recidivism more reliably than human judges do? Shouldn't policy be driven by science, and all that?
If they are demonstrated to work as well as or better than human judgment, why wouldn't you want to use them? ---------------------------------- "Go yell at an M&M"
So the AI overlords do, in fact, have your support, right?
So you didn't actually read the article? Or you didn't understand the objections to the use of this software tool? ---------------------------------- "Go yell at an M&M"
I would think that would have to be demonstrated before using it in the first place.
Nah. It's a mathematical model. All you need to do is get 97% of scientists to agree to it.
Again, only half joking.
What if the algorithms are demonstrated to work? What if they work as good or better than human judgement? What if they predict recidivism more reliably than human judges do? Shouldn't policy be driven by science, and all that?
Well one could always argue that additional non-algorithm factors come into play as well. Like punishment. So there will always be something at the judge's discretion.
I read the article. Is your only objection the fact that the algorithms are confidential?
That and the fact that these algorithms have been demonstrated to be worse than human predictions. ---------------------------------- "Go yell at an M&M"
I have no idea what, if any, point you are trying to make. What, exactly, do you think I'm objecting to? Do you think such objection is based solely on the content of that article? ---------------------------------- "Go yell at an M&M"
What, exactly, do you think I'm objecting to? Do you think such objection is based solely on the content of that article?
I think you're objecting to the fact that this company's algorithms are secret, and your contention that that its predictions are less reliable than human judges. Isn't that what you said?
If you're basing any of that off sources besides this article, feel free to share. I'm sure I'm not the only one in here who hasn't seen a comparison between the accuracy of predictions made by human judges and those made by AI algorithms.
"People think it must be fun to be a super genius, but they don't realize how hard it is to put up with all the idiots in the world."
What, exactly, do you think I'm objecting to? Do you think such objection is based solely on the content of that article?
I think you're objecting to the fact that this company's algorithms are secret, and your contention that that its predictions are less reliable than human judges. Isn't that what you said?
If you're basing any of that off sources besides this article, feel free to share. I'm sure I'm not the only one in here who hasn't seen a comparison between the accuracy of predictions made by human judges and those made by AI algorithms.
Weapons of Math Destruction! http://www.latimes.com/books/jacketcopy/la-ca-jc-cathy-oneil-20161229-story.html
I suspect that once we have Single Payer health care, the death panels will be similar computer systems.
i am working with one hand at the moment so can't copy paste the link to read. but agree with your conclusion. except they will be called life choice panels. ________ It doesn't really matter what Phil is saying, the music of his voice is the appropriate soundtrack for a bicycle race. HTupolev
Weapons of Math Destruction! http://www.latimes.com/books/jacketcopy/la-ca-jc-cathy-oneil-20161229-story.html
I suspect that once we have Single Payer health care, the death panels will be similar computer systems.
i am working with one hand at the moment so can't copy paste the link to read. but agree with your conclusion. except they will be called life choice panels.
They constantly try to escape from the darkness outside and within Dreaming of systems so perfect that no one will need to be good T.S. Eliot
What do scientists do when their models are flawed? They adjust the model, right?
Sounds like that particular model could be adjusted. I note there are several other models in use, and they might very well be more accurate. For that matter, while the model in your link appears flawed, it's still not clear that it's less accurate than human judges. The exact same complaints about racial disparities in sentencing have existed for decades.
"People think it must be fun to be a super genius, but they don't realize how hard it is to put up with all the idiots in the world."