<AR­TI­FI­CIAL IN­TEL­LI­GENCE

MICHAEL SPENCE AND FRED HU

The East African - - FRONT PAGE - Michael Spence, a No­bel lau­re­ate in eco­nomics, is pro­fes­sor of eco­nomics at New York Univer­sity. Fred Hu is chair­man and founder of Pri­mav­era Cap­i­tal Group, a Chin­abased global in­vest­ment firm.

We risk hav­ing dig­i­tal bor­ders as In­ter­net in­se­cu­rity fans abuse of data

The re­cent rev­e­la­tion that more than 50 mil­lion Face­book pro­files were har­vested by apps and given to po­lit­i­cal con­sul­tancy Cam­bridge An­a­lyt­ica has pro­duced a back­lash against the plat­form. But it is just the lat­est ex­am­ple of the risks as­so­ci­ated with the In­ter­net, which forms the core of to­day’s dig­i­tal rev­o­lu­tion.

Most of the dig­i­tal in­no­va­tions that have re­shaped the global econ­omy over the past 25 years rely on net­work con­nec­tiv­ity, which has trans­formed com­merce, com­mu­ni­ca­tion, ed­u­ca­tion and train­ing, sup­ply chains, and much more. Con­nec­tiv­ity also en­ables ac­cess to vast amounts of in­for­ma­tion, in­clud­ing in­for­ma­tion that un­der­pins ma­chine learn­ing, which is es­sen­tial to mod­ern ar­ti­fi­cial in­tel­li­gence.

Over the past 15 years or so, mo­bile In­ter­net has re­in­forced this trend, by rapidly in­creas­ing not just the num­ber of peo­ple who are con­nected to the In­ter­net, and thus able to par­tic­i­pate in the dig­i­tal econ­omy, but also the fre­quency and ease with which they can con­nect.

From GPS nav­i­ga­tion to ride-shar­ing plat­forms to mo­bile-pay­ment sys­tems, on-the-go con­nec­tiv­ity has had a far-reach­ing im­pact on peo­ple’s lives and liveli­hoods.

For years, it was widely be­lieved that an open In­ter­net – with stan­dard­ised pro­to­cols but few reg­u­la­tions – would nat­u­rally serve the best in­ter­ests of users, com­mu­ni­ties, coun­tries, and the global econ­omy. But ma­jor risks have emerged, in­clud­ing mo­nop­oly power for mega-plat­forms like Face­book and Google; vul­ner­a­bil­ity to at­tacks on crit­i­cal in­fra­struc­ture, in­clud­ing fi­nan­cial-mar­ket sys­tems and elec­toral pro­cesses; and threats to pri­vacy and the se­cu­rity of data and in­tel­lec­tual prop­erty.

Fun­da­men­tal ques­tions about the In­ter­net’s im­pact on po­lit­i­cal al­le­giance, so­cial co­he­sion, cit­i­zen aware­ness and en­gage­ment, and child­hood de­vel­op­ment also re­main.

As the In­ter­net and dig­i­tal tech­nolo­gies pen­e­trate economies and so­ci­eties more deeply, th­ese risks and vul­ner­a­bil­i­ties be­come in­creas­ingly acute. And, so far, the pre­dom­i­nant ap­proach to man­ag­ing them in the West — self- reg­u­la­tion by the com­pa­nies that pro­vide the ser­vices and own the data — does not seem to be work­ing. Ma­jor plat­forms can­not be ex­pected to re­move “ob­jec­tion­able” con­tent, for ex­am­ple, with­out guide­lines from reg­u­la­tors or courts.

Given this, we seem to be fac­ing a new tran­si­tion from the open In­ter­net of the past to one subject to more ex­ten­sive con­trol. But this process car­ries risks of its own.

Though there is a strong case for in­ter­na­tional co-op­er­a­tion, such an ap­proach seems un­likely in the cur­rent cli­mate of pro­tec­tion­ism and uni­lat­er­al­ism. It is not even clear that coun­tries will agree to treaties ban­ning cy­ber war­fare. Even if some sem­blance of in­ter­na­tional co-op­er­a­tion were mus­tered, non-state ac­tors would con­tinue to act as spoil­ers – or

worse. Against this back­ground, it seems likely that new reg­u­la­tions will be ini­ti­ated largely by in­di­vid­ual states, which will have to an­swer dif­fi­cult ques­tions. Who is re­spon­si­ble, and li­able, for data se­cu­rity? Should the state have ac­cess to user data, and for what pur­poses? Will users be al­lowed to main­tain anonymity online?

Coun­tries’ an­swers to such ques­tions will vary widely, ow­ing to fun­da­men­tal dif­fer­ences in their val­ues, prin­ci­ples, and gov­er­nance struc­tures. For ex­am­ple, in China, the au­thor­i­ties fil­ter con­tent deemed to be in­con­sis­tent with state in­ter­ests; in the West, by con­trast, there is no en­tity with le­git­i­mate au­thor­ity to fil­ter con­tent, ex­cept in ex­treme cases (for ex­am­ple, hate speech and child pornog­ra­phy). Even in ar­eas where there does seem to be some consensus – such as the un­ac­cept­abil­ity of dis­in­for­ma­tion or of for­eign med­dling in elec­toral pro­cesses – there is no agree­ment on the ap­pro­pri­ate rem­edy.

The lack of consensus or co-op­er­a­tion could lead to the emer­gence of na­tional dig­i­tal bor­ders, which would not only in­hibit flows of data and in­for­ma­tion, but also dis­rupt trade, sup­ply chains, and cross-bor­der in­vest­ment. Al­ready, most Usbased tech­nol­ogy plat­forms can­not op­er­ate in China, be­cause they can­not or will not ac­cept the au­thor­i­ties’ rules about state ac­cess to data and con­trol over con­tent.

Mean­while, the US has blocked the Chi­nese com­pany Huawei from in­vest­ing in soft­ware star­tups, pro­vid­ing net­work equip­ment to wire­less car­ri­ers, or (along with ZTE) sell­ing mo­bile phones in the US mar­ket, ow­ing to the firm’s al­leged ties to the Chi­nese govern­ment. Huawei and ZTE both main­tain that their ac­tiv­i­ties are purely com­mer­cial, and that they play by the rules wher­ever they op­er­ate, but US of­fi­cials con­tinue to in­sist that the com­pa­nies pose a se­cu­rity risk. By con­trast, nearly all Euro­pean coun­tries, in­clud­ing the United King­dom, are re­cep­tive to Huawei and ZTE, both of which are ma­jor play­ers in Europe. Yet Europe is cre­at­ing its own bar­ri­ers, with new data-pro­tec­tion and pri­vacy rules that may well im­pede the ap­pli­ca­tion of ma­chine learn­ing. Un­like China and the US, Europe is not yet home to a mega plat­form of the type that is lead­ing the way in ma­chine-learn­ing in­no­va­tions.

With the en­tire global econ­omy be­com­ing in­ex­tri­ca­bly linked to the In­ter­net and dig­i­tal tech­nolo­gies, stronger reg­u­la­tion is more im­por­tant than ever. But if that reg­u­la­tion is frag­mented, clumsy, heavy-handed, or in­con­sis­tent, the con­se­quences for eco­nomic in­te­gra­tion – and, in turn, pros­per­ity – could be se­vere.

Be­fore the world adopts in­ef­fec­tive or coun­ter­pro­duc­tive so­lu­tions, pol­i­cy­mak­ers should think care­fully about how best to ap­proach reg­u­la­tion. If we can­not agree on ev­ery de­tail, per­haps we can at least iden­tify a set of shared prin­ci­ples that can form the ba­sis of mul­ti­lat­eral agree­ments that pro­scribe de­struc­tive ac­tiv­ity such as the mis­use of data, thereby help­ing to pre­serve an open global econ­omy.

For years, it was widely be­lieved that an open In­ter­net – with stan­dard­ised pro­to­cols but few reg­u­la­tions – would nat­u­rally serve the best in­ter­ests of users, com­mu­ni­ties, coun­tries, and the global econ­omy

Newspapers in English

Newspapers from Kenya

© PressReader. All rights reserved.