Pretoria News - - Focus - PRO­FES­SOR LOUIS FOURIE

THE HEAD of the In­ter­na­tional Mone­tary Fund yes­ter­day de­fended Fed­eral Re­serve chair­per­son Jerome Pow­ell one day af­ter US Pres­i­dent Don­ald Trump crit­i­cised the Amer­i­can cen­tral bank, which has been rais­ing US in­ter­est rates, as “crazy”. “I wouldn’t as­so­ciate Jay Pow­ell with crazi­ness,” manag­ing di­rec­tor Chris­tine La­garde told CNBC in an in­ter­view. “He comes across – and mem­bers of his board – as ex­tremely se­ri­ous, solid and cer­tainly keen to base their de­ci­sions on ac­tual in­for­ma­tion.” Trump made the com­ment in re­sponse to Wall Street’s sell-off af­ter the S&P 500 and the Dow marked their worst losses in eight months. The slump trig­gered a surge of global sell­ing that sent Euro­pean stocks to a more than an 18-month low yes­ter­day and knocked down Asian shares. Wor­ries about higher bor­row­ing costs have sparked the US stock sales in re­cent days, while a spike in US Trea­sury yields and solid US eco­nomic data have raised con­cerns that the Fed may pick up the pace of its in­ter­est rate hikes. “Ac­tu­ally, it’s a cor­rec­tion we’ve been wait­ing for a long time, but I re­ally dis­agree with what the Fed is do­ing,” Trump said. “I think the Fed is mak­ing a mis­take. They’re so tight. I think the Fed has gone crazy.” Trump tapped Pow­ell to lead the bank but has pub­licly crit­i­cised the Fed un­der his ten­ure. | Reuters THE COM­BI­NA­TION of Ar­ti­fi­cial In­tel­li­gence (AI), in par­tic­u­lar ma­chine learn­ing, deep learn­ing and cog­ni­tive com­put­ing, and ro­bot­ics to de­liver ar­ti­fi­cially in­tel­li­gent ro­bots are trans­fig­ur­ing the way we live and work.

How­ever, not all ro­bots are ar­ti­fi­cially in­tel­li­gent. Un­til re­cently, most in­dus­trial ro­bots were rather limited in their func­tion­al­ity and could only be pro­grammed to carry out a re­cur­ring se­quence of move­ments. But repet­i­tive move­ments don’t re­quire AI. If the ro­bot is ex­pected to un­der­take more com­plex tasks, AI al­go­rithms and ma­chine learn­ing logic are nec­es­sary.

AI is not merely about ma­chines think­ing for them­selves, but en­tails a com­plex set of in­puts and out­puts that al­lows the ma­chine to make smart de­ci­sions and re­spond to its en­vi­ron­ment in real time based on a deep knowl­edge base, as well as the abil­ity to con­stantly add to this knowl­edge through learn­ing.

Per­haps the best-known ar­ti­fi­cially in­tel­li­gent ro­bot is Sophia, a hu­manoid ro­bot mod­elled af­ter ac­tress Au­drey Hep­burn.

Sophia is known for her amaz­ing hu­man-like ap­pear­ance and be­hav­iour and can dis­play a range of hu­man ges­tures and more than 50 fa­cial ex­pres­sions. She pos­sesses three dis­tinc­tively hu­man fea­tures, namely, cre­ativ­ity, em­pa­thy and com­pas­sion.

Among other tech­nolo­gies, Sofia uses AI, vis­ual data pro­cess­ing, voice recog­ni­tion and fa­cial recog­ni­tion and is de­signed to get smarter over time. This en­ables Sophia to fol­low faces, recog­nise in­di­vid­u­als, sus­tain eye con­tact, an­swer cer­tain ques­tions, have sim­ple con­ver­sa­tions and walk around pur­pose­fully.

Sophia has been in­ter­viewed on nu­mer­ous shows in the same man­ner as hu­mans. Some an­swers did not make sense, while oth­ers have im­pressed in­ter­view­ers.

She has ap­peared on stage as a panel mem­ber and pre­sen­ter in high­level con­fer­ences, dis­cussing how in­tel­li­gent ro­bots will play an in­creas­ing role in peo­ple’s lives.

The man­u­fac­tur­ers be­lieve that Sophia will even­tu­ally be able to serve in health and el­derly care, cus­tomer ser­vice, ther­apy and ed­u­ca­tion.

But a se­ri­ous ques­tion arises: Does Sophia have hu­man-equiv­a­lent in­tel­li­gence or is she just a chat­bot with a face? Al­though re­mark­able progress has been made, Sophia still has a long way to go in terms of ar­ti­fi­cial gen­eral in­tel­li­gence and a sen­tient be­ing.

What is cer­tain is that we will see many more ar­ti­fi­cially in­tel­li­gent ro­bots in our homes.

The first home ro­bots ap­peared in the 1990s and have as­sisted with many do­mes­tic tasks such as clean­ing, en­ter­tain­ment and do­mes­tic se­cu­rity. How­ever, the use of more ad­vanced AI tech­nolo­gies in home ro­bots is rel­a­tively new.

The new­est clean­ing ro­bots driven by AI have ad­vanced de­ci­sion-mak­ing ca­pa­bil­i­ties to recog­nise ob­sta­cles and com­pute the most ef­fi­cient route; speech recog­ni­tion to take user com­mands and re­port their cur­rent sta­tus; and in­ter­ac­tive draw­ing ca­pa­bil­i­ties to draw maps of their en­vi­ron­ment.

In en­ter­tain­ment, ar­ti­fi­cial in­tel­li­gent ro­bots, us­ing deep AI and com­puter vi­sion, are in­creas­ingly dis­play­ing inim­itable per­son­al­i­ties and hu­man emo­tions. They can iden­tify and re­mem­ber peo­ple and have “evolv­ing” per­son­al­i­ties to adapt to their own­ers.

Fur­ther­more, they can in­ter­pret users’ fa­cial ex­pres­sions, vo­cal in­to­na­tions and lin­guis­tic pat­terns; proac­tively start con­ver­sa­tions rather than re­spond­ing to users’ com­mands; and sim­u­late emo­tions like em­pa­thy. These ro­bots also have the abil­ity to re­trieve in­for­ma­tion and com­mu­ni­cate with other con­nected de­vices and, for in­stance, would au­to­mat­i­cally ad­just the mu­sic ac­cord­ing to the mood of a per­son.

Ar­ti­fi­cially in­tel­li­gent ro­bots are also be­com­ing more com­mon in home se­cu­rity and surveil­lance. They can recog­nise faces and de­tect sus­pi­cious sounds, so that it can warn the home­owner of in­trud­ers. Some can even pre­dict and dis­rupt crimes be­fore they oc­cur.

In the com­ing years, the need for med­i­cal re­ha­bil­i­ta­tion will in­crease as the rate of sur­vival af­ter dis­eases with se­vere func­tional lim­i­ta­tions, such as a stroke, will in­crease. So­cially As­sis­tive Ro­bots are al­ready be­ing used in re­ha­bil­i­ta­tion. A mile­stone was reached when the first ar­ti­fi­cially in­tel­li­gent ro­bot in space, named Ci­mon, ar­rived at the In­ter­na­tional Space Sta­tion on July 2, 2018. It uses IBM’s fa­mous Wat­son sys­tem and can com­mu­ni­cate with peo­ple, know­ing whom it’s talk­ing to due to fa­cial-recog­ni­tion soft­ware.

Ci­mon’s pur­pose is to as­sist as­tro­naut Alexan­der Gerst in sev­eral in­ves­ti­ga­tions by ac­cess­ing a large amount of rel­e­vant in­for­ma­tion, in­clud­ing pho­tos and videos.

Ci­mon is smart enough to deal with ques­tions be­yond the in­ves­ti­ga­tions that Gerst might have.

Ci­mon is an ex­per­i­ment of hu­man-ma­chine in­ter­ac­tion and col­lab­o­ra­tion in space with the aim of even­tu­ally help­ing as­tro­nauts re­pair dam­aged space­craft sys­tems or treat­ing sick crew mem­bers.

Ar­ti­fi­cially in­tel­li­gent ro­bots cre­ate won­der­ful op­por­tu­ni­ties, but may also be fear-in­duc­ing. For many peo­ple, the idea of in­tel­li­gent ro­bots cre­ates vi­sions of ro­bots that take over the world.

How­ever, the chance that ro­bots want to rise up and dom­i­nate Earth are as slim as that they want to drink al­co­hol or take drugs. That is uniquely hu­man.

But we will still have to en­sure that AI and ro­bot­ics are de­vel­oped con­sci­en­tiously and con­strained by moral stan­dards and ethics.

Newspapers in English

Newspapers from South Africa

© PressReader. All rights reserved.