Page 102 - 202

ɛ˖ၾٟึ߅ኪᔊৃc20 ՜ 2 ಂg108 ϋ 3 ˜

by AIjɛٟνОΪᏐ AI ࠠอܠϽ޼Ӻ˙όkɛٟνОϞп׵ AI €HSS for
AIjɛٟνО˸ІԒዹतᝈᓃઞী AI ٙঐၾʔঐd࿁ AI ೯࢝౤Զ্ᘠk
2
ତϞྼਕޢ࿁ AI ٙɛٟᅂᚤʊϞᙃʿf Ͼ AI ٙɛٟจ଄̙ਗ਼ AI ൖމอ
3
ٙᔼᐕਖ਼ุҦঐe᜗Փe༟๕e˖ʷٙڢɛБਗ٫ €nonhuman actor d͟ᔼᐕ
ٟึኪᙄᕚ˴ൿ஼ɓઞীfίϤೌج୚ሔf

ɚe ɛٟνОԴ͜ AI

һආɓӉdνΝཥ໘࿁୕ࠇ˙جٙᅂᚤdAI ɰਗ਼ϓމࠠࠅٙɛٟ޼Ӻʈ
ՈfԷνdМᎇيᑌၣٙ༟ࣘႎණe˖οၾႧࠪ AI ፫ᗆeଉܓኪ୦စၑجഃڮ
ϓٙɽᅰኽ €big data ޼Ӻ˙جdঐ߉ॎ΋ۃɛɢيɢࠢՓdІਗʷʱٜؓટ͟

͎᜗ႎණٙᔼᐕၾ਄ੰᅰኽࠠ࿴ඎʷٟึତྼfШ࿁ɛٟԸႭdவ၇อϓఱɰ
ਗ਼੭Ըڢ AI הঐ༆Ӕًٙ࿒d̙ঐᙄᕚܼ̍j
1. ʫᅰኽ €infra-dataj͟΢၇ช಻ኜeᅰЗ̻ၽאيᑌၣٙᅰኽ͛ପԨڢѩሯ

˲੭Ϟ΢၇จྡdˈՉ݊຅ලଷʷ€black-boxedٙ AI ʊ຾ʧɝਠุၾ݁ط
4
ٙ༟ࣘପ͛eᝈ಻eરҏၾႎණཀ೻ࣛdਗ਼ˏ೯ٙอÿڢཫಂܝ؈d ຅வԬ
༟ࣘɦϓމ AI ኪ୦ٙਿᓾdһ̙ঐႬኬҢࡁ࿁ତྼٙଣ༆f 5
2. ێᅰኽ €thick dataj࿁׵ᅰኽπίၾପ͛ཀ೻ٙএഖࠠ࿴dɓٜ݊ν͏ૄႦ
אྼЪ޼Ӻഃ͞௉ሜݟٙ੶ධdϾவ၇࿁ٟึ͛ݺএഖආБଉ౜ €thick

descriptionٙ˙جdϞп׵ί஗ѩሯʷٙᅰኽऎd༕ᙑ΢၇ྼЪeਜЗၾઋ
6
ྤٙএഖϾһމ൨ڐٟึତྼf வʊ຾ίɓධࡲఁٙ͏ૄႦ՘Ъ̻ၽක֐ྼ
᜕f 7

3. ܝᅰኽ €meta-datajɽᅰኽٙΎ͛ପཀ೻ €Էν͜ᅰኽ৅ᇖ AI စၑج Ԩʔ
ɓ֛ቇ຅ˀ݈ତྼdϾ̙ঐˏ೯ʔᓙ჈ᓥٙᅰኽႬࢨ৙ਸ਼fԷν˸ତϞ༟ࣘ

2
਷̮ሗԈ IBM ٙ Watson ӻ୕d਷ʫሗԈҽʾਖ਼ €2018AI ᔼᐕɽ͊Ըj̨ᝄୋɓ͉౽ᅆᔼᐕᗫᒟ
జѓ‘dอ̹̏jλɛ̈وf
3
؍˖๕ €2018f”ᔼᐕٙ݁ط׌j੽ٟึeٝᗆՑ͉᜗ሞ݁طၾ͉ήᔼᐕྼЪ͉ٙ᜗ሞ݁ط޼Ӻᕚ
ၤ•d߅Ҧeᔼᐕၾٟึ‘ 26 ಂdࠫ 115-184f
4
Frank Pasquale. (2016). The Black Box Society: The Secret Algorithms that Control Money and Information.
Harvard University Press.
5
Nature, ‘AI can be sexist and racist — it’s time to make it fair’. 18 JULY 2018.
6
Ethnography Matters, ‘Big Data Needs Thick Data’ May 13, 2013.
7
Ethnography Matters, ‘Innovation in Asthma Research: Using Ethnography to Study a Global Health
Problem’ ӻΐf
100