‘There’s a layer of anx­i­ety in all we do’

Big tech firms must be reg­u­lated be­fore it’s too late, says Joseph Stiglitz

The Guardian Weekly - - Finance | Interview - Ian Sam­ple

It must be hard for Joseph Stiglitz to re­main an op­ti­mist in the face of the grim fu­ture he fears may be com­ing. The No­bel lau­re­ate and for­mer chief economist at the World Bank has thought care­fully about how ar­ti­fi­cial in­tel­li­gence will af­fect our lives. On the back of the technology, we could build our­selves a richer so­ci­ety and per­haps en­joy a shorter work­ing week, he says. But there are count­less pit­falls to avoid on the way. The ones Stiglitz has in mind are hardly triv­ial. He wor­ries about ham­fisted moves that lead to rou­tine ex­ploita­tion in our daily lives, that leave so­ci­ety more di­vided than ever and threaten the fun­da­men­tals of democ­racy.

“Ar­ti­fi­cial in­tel­li­gence and robo­ti­sa­tion have the po­ten­tial to in­crease the pro­duc­tiv­ity of the econ­omy and, in prin­ci­ple, that could make ev­ery­body bet­ter off,” he says. “But only if they are well man­aged.”

The Columbia Uni­ver­sity pro­fes­sor this week de­liv­ered the lat­est lec­ture in the Royal So­ci­ety’s You and AI se­ries in Lon­don. Stiglitz talked about the fu­ture of work, an area where pre­dic­tions have been con­tra­dic­tory and un­nerv­ing. Last month, the Bank of Eng­land’s chief economist, Andy Hal­dane, warned that “large swathes” of Bri­tain’s work­force face un­em­ploy­ment as AI and other tech­nolo­gies au­to­mate more jobs. He had less to say about the new po­si­tions AI may cre­ate. A re­port from Price­wa­ter­house­C­oop­ers in July ar­gued that AI may cre­ate as many jobs as it de­stroys – per­haps even more. As with the in­dus­trial rev­o­lu­tion, the mis­ery would come not from a lack of work, but the difficulty in switch­ing from one job to an­other.

A dis­tinc­tion Stiglitz makes is be­tween AI that re­places work­ers and AI that helps peo­ple to do their jobs bet­ter. It al­ready helps doc­tors to work more ef­fi­ciently. But well-trained AIs are now bet­ter at spot­ting breast tu­mours and other cancers than ra­di­ol­o­gists. Does that mean un­em­ploy­ment for ra­di­ol­o­gists? It is not so straight­for­ward, says Stiglitz. “Read­ing an MRI scan is only part of the job that per­son per­forms, but you can’t eas­ily sep­a­rate that task from the oth­ers.”

Stiglitz won the No­bel prize for eco­nom­ics in 2001 for his analy­ses of im­per­fect in­for­ma­tion in mar­kets. A year later, he pub­lished Glob­al­i­sa­tion and Its Dis­con­tents, a book that laid bare his dis­il­lu­sion with the In­ter­na­tional Mon­e­tary Fund and, by ex­ten­sion, the US Trea­sury. Trade ne­go­ti­a­tions, he ar­gued, were driven by multi­na­tion­als at the ex­pense of work­ers and or­di­nary cit­i­zens. “What I want to em­pha­sise is that it is time to fo­cus on the pub­lic-pol­icy is­sues sur­round­ing AI, be­cause the con­cerns are a con­tin­u­a­tion of the con­cerns that glob­al­i­sa­tion and in­no­va­tion have brought us. We were slow to grasp what they were do­ing and we shouldn’t make that mis­take again.”

Be­yond the im­pact of AI on work, Stiglitz sees more in­sid­i­ous forces at play. “In your in­ter­ac­tions with Google, Facebook, Twit­ter and oth­ers, they gather an aw­ful lot of data about you. If that data is com­bined with other data, then com­pa­nies have a great deal of in­for­ma­tion about you as an in­di­vid­ual – more in­for­ma­tion than you have on yourself,” he says.

“They know, for ex­am­ple, that peo­ple who search this way are will­ing to pay more. They know ev­ery store you’ve vis­ited. That means that life is go­ing to be in­creas­ingly un­pleas­ant, be­cause your de­ci­sion to shop in a cer­tain store may re­sult in you pay­ing more money. To the ex­tent that peo­ple are aware of this game, it dis­torts their be­hav­iour. What is clear is that it in­tro­duces a level of anx­i­ety in ev­ery­thing we do and it in­creases in­equal­ity even more.”

Stiglitz poses a ques­tion that he sus­pects tech firms have faced in­ter­nally. “Which is the eas­ier way to make a buck: fig­ur­ing out a bet­ter way to ex­ploit some­body, or mak­ing a bet­ter prod­uct? With the new AI, it looks like the an­swer is find­ing a bet­ter way to ex­ploit some­body.”

Rev­e­la­tions about how Rus­sia turned to Facebook, Twit­ter and Google to in­ter­fere with the 2016 US elec­tion brought home how ef­fec­tively peo­ple can be tar­geted. In the US in par­tic­u­lar, there has been a will­ing­ness to leave tech firms to thrash out de­cent rules of be­hav­iour and ad­here to them, Stiglitz be­lieves. One of the many rea­sons is that the com­plex­ity of the technology can make it in­tim­i­dat­ing. “It over­whelms a lot of peo­ple and their re­sponse is: ‘We can’t do it, the gov­ern­ment can’t do it, we have to leave it to the tech gi­ants.’”

But Stiglitz thinks that view is chang­ing. The mea­sures Stiglitz pro­poses are broad and it is hard to see how they could be brought in swiftly. You can’t al­low the tech gi­ants to do it,” he says. “It has to be done pub­licly with an aware­ness of the dan­ger that the tech firms rep­re­sent.”

Fresh poli­cies are needed to curb monopoly pow­ers and re­dis­tribute the wealth that is con­cen­trated in the lead­ing AI firms, he adds. This month, Ama­zon be­came the sec­ond com­pany, af­ter Ap­ple, to reach a mar­ket val­u­a­tion of $1tn. The pair are now worth more than the top 10 oil com­pa­nies com­bined. “When you have so much wealth con­cen­trated in the hands of rel­a­tively few, you have a more un­equal so­ci­ety and that is bad for our democ­racy,” says Stiglitz.

Taxes are not enough. To Stiglitz, this is about labour bar­gain­ing power, in­tel­lec­tual prop­erty rights, re­defin­ing and en­forc­ing com­pe­ti­tion laws, cor­po­rate gov­er­nance laws and the way the fi­nan­cial sys­tem op­er­ates. “It’s a much broader agenda than just re­dis­tri­bu­tion,” he says.

He is not a fan of uni­ver­sal basic in­come. Ad­vo­cates ar­gue that, as tech firms gather ever more wealth, UBI could help to re­dis­tribute the pro­ceeds and en­sure that ev­ery­one ben­e­fits. But Stiglitz does not be­lieve it is what most peo­ple want.

“If we don’t change our over­all eco­nomic and pol­icy frame­work, what we’re go­ing to­wards is greater wage in­equal­ity, greater in­come and wealth in­equal­ity and prob­a­bly more un­em­ploy­ment and a more di­vided so­ci­ety. But none of this is in­evitable,” he says. “By chang­ing the rules, we could wind up with a richer so­ci­ety, with the fruits more equally di­vided, and quite pos­si­bly where peo­ple have a shorter work­ing week. We’ve gone from a 60-hour work­ing week to a 45-hour week and we could go to 30 or 25.”

None of this will hap­pen overnight, he warns. A more ro­bust pub­lic de­bate around AI and work is needed to throw up new ideas, for a start. “Sil­i­con Val­ley may hire a dis­pro­por­tion­ate frac­tion [of peo­ple who work in AI], but it may not take that many peo­ple to fig­ure it out, in­clud­ing peo­ple from Sil­i­con Val­ley who have be­come dis­grun­tled with what has been go­ing on,” he says. “Peo­ple will, and have al­ready be­gun to, think about new ideas. There will be peo­ple with skills who try to work out so­lu­tions.”

‘With new rules, we could have a richer so­ci­ety – the fruits more equally di­vided’

Newspapers in English

Newspapers from UK

© PressReader. All rights reserved.