Call for AI-spe­cific hu­man rights frame­work

The Irish Times - Business - - TECHNOLOGY - Marie Bo­ran ac­cess­now.org/map­ping-ar­ti­fi­cial­in­tel­li­gence-strate­gies-in-europe/

An oft-used ar­gu­ment against govern­ment reg­u­la­tion of emerg­ing tech­nolo­gies is the sti­fling of in­no­va­tion.

On the other side of this ar­gu­ment is the need to pro­vide a frame­work within which cit­i­zens can be pro­tected from threats to pri­vacy, au­ton­omy, well­be­ing and other as­pects of hu­man rights that may be af­fected, as tech­nolo­gies such as ar­ti­fi­cial in­tel­li­gence are in­creas­ingly in­cor­po­rated into ev­ery­thing from ed­u­ca­tion and trans­port to health and fi­nance.

In­ter­na­tional non-profit dig­i­tal rights group Ac­cess Now has pub­lished a re­port assess­ing ex­ist­ing pan-Euro­pean strate­gies for AI de­vel­op­ment and reg­u­la­tion with sev­eral rec­om­men­da­tions, in­clud­ing a strat­egy of trans­parency: “Pub­lic bod­ies should en­gage the pub­lic in these de­ci­sions [on AI] to the fullest pos­si­ble ex­tent.”

The re­port also noted that “sur­pris­ingly lit­tle” work has been done by Euro­pean gov­ern­ments to ap­ply ex­ist­ing anti-dis­crim­i­na­tion law to AI sys­tems, which is needed in light of al­go­rith­mic bias such as job ad­ver­tise­ments on Face­book aimed at men rather than women and crim­i­nal jus­tice soft­ware cal­cu­lat­ing re-of­fend­ing rates based on eth­nic back­ground.

Ac­cess Now also said an area of con­cern is “strate­gies which, via pay­ing a nod to ‘ethics’, mainly ex­press will­ing­ness to loosen the reg­u­la­tory en­vi­ron­ment. Au­thor­i­ties should be vig­i­lant that ethics do not be­come a smoke­screen for an un­reg­u­lated tech­ni­cal en­vi­ron­ment.”

Newspapers in English

Newspapers from Ireland

© PressReader. All rights reserved.