How would you like to be sentenced by a computer program?

quote

when chief justice john g. roberts jr. visited rensselaer polytechnic institute last month, he asked startling question, 1 overtones of science fiction.

“can foresee day,” asked shirley ann jackson, president of college in upstate new york, “when smart machines, driven artificial intelligences, assist courtroom fact-finding or, more controversially even, judicial decision-making?”

chief justice’s answer more surprising question. “it’s day that’s here,” said, “and it’s putting significant strain on how judiciary goes doing things.”

may have been thinking case of wisconsin man, eric l. loomis, who sentenced 6 years in prison based in part on private company’s proprietary software. mr. loomis says right due process violated judge’s consideration of report generated software’s secret algorithm, 1 mr. loomis unable inspect or challenge.

click expand...

https://mobile.nytimes.com/2017/05/...by-a-software-programs-secret-algorithms.html

maybe see nothing wrong this? @ minimum defendant has constitutional right inspect , challenge all evidence used against him. maybe think big data unbiased , fair?

perhaps should consider this:

quote

in debates on future of artificial intelligence, many experts think of new systems coldly logical , objectively rational. in new study, researchers have demonstrated how machines can reflections of us, creators, in potentially problematic ways. common machine learning programs, when trained ordinary human language available online, can acquire cultural biases embedded in patterns of wording, researchers found. these biases range morally neutral, preference flowers on insects, objectionable views of race , gender

click expand...

https://www.sciencedaily.com/releases/2017/04/170413141055.htm

quote

machine learning means derive artificial intelligence discovering patterns in existing data. here, show applying machine learning ordinary human language results in human-like semantic biases. replicated spectrum of known biases, measured implicit association test, using used, purely statistical machine-learning model trained on standard corpus of text world wide web. our results indicate text corpora contain recoverable , accurate imprints of our historic biases, whether morally neutral toward insects or flowers, problematic toward race or gender, or veridical, reflecting status quo distribution of gender respect careers or first names. our methods hold promise identifying , addressing sources of bias in culture, including technology.

click expand...

http://science.sciencemag.org/content/356/6334/183

luck.

 

that's not quite happened. man convicted assuming jury(good luck finding info via google original trial), , sentence handed down judge. 6 years pretty light compared fact that, a. he's been convicted of sexual assault. in stolen car , evaded leo. history of violence , grand theft. care meter pegged @ zero.

 

Forums Mac Community Politics, Religion, Social Issues

• iPhone
• Mac OS & System Software
• iPad
• Apple Watch
• Notebooks
• iTunes
• Apple ID
• iCloud
• Desktop Computers
• Apple Music
• Professional Applications
• iPod
• iWork
• Apple TV
• iLife
• Wireless