Poročilo o avtomatiziranju družbe: slovenski primer / Automating Society report: Slovenian case 

Človeške družbe najverjetneje ne bodo uničili superinteligentni morilski roboti, ki pogosto nastopajo v filmih in serijah. Takšne podobe inteligentnih strojev so sicer zanimive, vendar hkrati odvračajo pozornost od bistveno bolj verjetnih načinov uporabe strojnega učenja, ki jih že uvajajo številne države EU. Algoritmi že marsikje odločajo o odpuščanju in zaposlitvah delavcev, podeljujejo ali jemljejo socialno pomoč, izračunavajo tveganje pri podeljevanju bančnih kreditov in iščejo nevarne voznike ali otroke z učnimi težavami. Toda njihove odločitve niso nevtralne, saj so tudi v programske vrstice vpisane vse značilnosti in ideologije sodobnih družb, kar lahko še povečuje diskriminacijo in družbeno neenakost. Pogosto v imenu večje učinkovitosti in obljube prihrankov javnega denarja.

It is unlikely that human societies will be destroyed by superintelligent killer robots often featured in movies and series. Such images of intelligent machines are interesting, however, they divert attention away from the far more likely uses of machine learning that had been already introduced by many EU countries. Algorithms decide whether to dismiss or hire workers, grant or refuse to provide welfare benefits, calculate bank credit risk, discover dangerous drivers or children with learning difficulties. Decisions made by algorithms are not neutral, because the computer code includes all the characteristics and ideologies of modern societies, which can further increase discrimination and social inequality. Often in the name of greater efficiency and the promise of saving public money

festivalgrounded@gmail.com