Ro­bots spark charms and fears

Shanghai Daily - - TECHNOLOGY - Clara Wright

Sport­ing a trendy brown bob, a hu­manoid ro­bot named Erica chats to a man in front of stunned au­di­ence mem­bers in Madrid. She and oth­ers like her are a prime fo­cus of ro­botic re­search, as their un­canny hu­man form could be key to in­te­grat­ing such ma­chines into our lives, said re­searchers gath­ered last week at the an­nual In­ter­na­tional Con­fer­ence on In­tel­li­gent Ro­bots.

“You men­tioned pro­ject man­age­ment. Can you please tell me more?” Erica, who is play­ing the role of an em­ployer, asks the man.

She may not un­der­stand the con­ver­sa­tion, but she’s been trained to de­tect key words and re­spond to them.

A source of con­tro­versy due in part to fears for hu­man em­ploy­ment, the pres­ence of ro­bots in our daily lives is nev­er­the­less in­evitable, en­gi­neers at the con­fer­ence said.

The trick to mak­ing them more palat­able, they added, is to make them look and act more hu­man so that we ac­cept them into our lives more eas­ily.

In ag­ing so­ci­eties, “ro­bots will co­ex­ist with hu­mans sooner or later,” said Hiroko Kamide, a Ja­pa­nese psy­chol­o­gist who spe­cial­izes in re­la­tions be­tween hu­mans and ro­bots.

Wel­com­ing ro­bots into house­holds or work­places in­volves de­vel­op­ing “mul­ti­pur­pose ma­chines that are ca­pa­ble of in­ter­act­ing” with hu­mans with­out be­ing dan­ger­ous, said Philippe Soueres, head of the ro­bot­ics de­part­ment at a lab­o­ra­tory be­long­ing to France’s CNRS sci­en­tific in­sti­tute.

As such, ro­bots must move around “in a sup­ple way” de­spite their rigid me­chan­ics and stop what they are do­ing in case of any un­fore­seen event, he added.

That’s why peo­ple are choos­ing “mod­u­lar sys­tems shaped like hu­man bod­ies” which are meant to eas­ily fit into re­al­world en­vi­ron­ments built for hu­mans.

For in­stance At­las, a hu­manoid ro­bot made by Bos­ton Dy­nam­ics, can run on dif­fer­ent types of sur­faces.

In Madrid, Marc Raib­ert, founder of the US firm, played a video show­ing At­las do­ing a back­flip.

In a sign of fears over the po­ten­tial fu­ture uses for th­ese hu­manoids, Amnesty In­ter­na­tional has ac­cused At­las, fi­nanced by an agency of the US De­part­ment of De­fense, of be­ing a “killer ro­bot” made for fu­ture war­fare.

An­other ex­am­ple of hu­manoids pre­sented in Madrid is Ta­los, a ro­bot made by Span­ish com­pany Pal Ro­bot­ics shown test­ing his sta­bil­ity on a bal­ance board.

While it may not be the only form used for those com­ing into con­tact with hu­mans, “it’s eas­ier for peo­ple to ac­cept the ro­bots when they have hu­man-like faces be­cause peo­ple can ex­pect how the ro­bots will move, will re­act,” said Kamide. That’s com­fort­ing, but it also has its lim­its.

Ja­pa­nese re­searcher Masahiro Mori’s “un­canny val­ley” the­ory, which he de­vel­oped in the 1970s, states that we re­act pos­i­tively to ro­bots if they have phys­i­cal fea­tures fa­mil­iar to us but they dis­turb us if they start look­ing too much like us.

“You can’t ever make a per­fect hu­man face” and this im­per­fec­tion pro­vokes a feel­ing of “re­jec­tion” among hu­mans, said Miguel Salichs, a pro­fes­sor at the ro­bot­ics lab of Madrid’s Car­los III Uni­ver­sity. As such, he chose to fash­ion his ro­bot Mini Mag­gie into a small car­toon an­i­mal.

In Ja­pan, ro­bots like Erica are al­ready used as re­cep­tion­ists.

But for one of their mak­ers, Hiroshi Ishig­uro, a pro­fes­sor at Osaka Uni­ver­sity, hu­manoids are above all “a very im­por­tant tool to un­der­stand hu­mans.”

Re­searchers have to think hard about the hu­man form and how hu­mans in­ter­act to de­velop ro­bots that look like them. “We un­der­stand the hu­mans by us­ing ro­bots, the im­por­tance for ex­am­ple of eye gaz­ing,” said Ishig­uro, who has also made ro­bots that look like dead celebri­ties, or “mov­ing stat­ues.”

He be­lieves that hu­manoids are best to im­prove in­ter­ac­tions be­tween ro­bots and hu­mans.

“The hu­man brain that we have has many func­tions to rec­og­nize hu­mans. The nat­u­ral in­ter­face for the hu­mans is the hu­mans,” said Ishig­uro.

For Jur­gen Sch­mid­hu­ber, pres­i­dent of ar­ti­fi­cial in­tel­li­gence start-up NNAISENSE, ro­bots — hu­manoid or not — will be part of our fu­ture. They won’t just im­i­tate hu­mans but will solve prob­lems by ex­per­i­ment­ing them­selves thanks to ar­ti­fi­cial in­tel­li­gence with­out “a hu­man teacher,” he be­lieves. Sit­ting on her chair, Erica nods her head.

Newspapers in English

Newspapers from China

© PressReader. All rights reserved.