Rise of the ma­chines must be mon­i­tored

AI ac­tiv­ity in banks could drive prof­itabil­ity: FSB

Global Times - Weekend - - FRONT PAGE -

Re­plac­ing bank and in­sur­ance work­ers with ma­chines risks creat­ing a de­pen­dency on out­side tech­nol­ogy com­pa­nies be­yond the reach of reg­u­la­tors, the global Fi­nan­cial Sta­bil­ity Board (FSB) said on Wed­nes­day.

The FSB, which co­or­di­nates fi­nan­cial reg­u­la­tion across the Group of 20 economies (G20), said in its first re­port on ar­ti­fi­cial in­tel­li­gence (AI) and ma­chine learn­ing that the risks they pose need mon­i­tor­ing.

AI and ma­chine learn­ing re­fer to tech­nol­ogy that is re­plac­ing tra­di­tional meth­ods to as­sess the cred­it­wor­thi­ness of cus­tomers and to crunch data, price in­sur­ance contracts and spot prof­itable trades across mar­kets.

There are no in­ter­na­tional reg­u­la­tory stan­dards for AI and ma­chine learn­ing, but the FSB has left open whether new rules are needed.

Data on rapidly grow­ing us­age of AI is largely un­avail­able, leav­ing reg­u­la­tors un­sure about the im­pact of po­ten­tially new and un­ex­pected links be­tween mar­kets and banks, the re­port said.

AI could, for ex­am­ple, lead to “non­sus­tain­able” in­creases in credit by au­tomat­ing credit scor­ing.

While AI shows sub­stan­tial prom­ises if risks are prop­erly man­aged, it could cre­ate too much de­pen­dency among banks and in­sur­ers on the few spe­cial­ist busi­nesses that pro­vide AI tech­nol­ogy.

Ex­pected rapid growth in AI also raises the prospect of out­side tech­nol­ogy play­ers ex­pand­ing their in­flu­ence over the fi­nance sec­tor.

“This could in turn lead to the emer­gence of new sys­tem­i­cally im­por- tant play­ers that could fall out­side the reg­u­la­tory perime­ter,” the FSB said.

If a ma­jor AI provider went bust, it could lead to op­er­a­tional dis­rup­tions at a large num­ber of fi­nan­cial firms at the same time, es­pe­cially if used in “mis­sion crit­i­cal” ap­pli­ca­tions, the re­port said.

Reg­u­la­tors could also find it dif­fi­cult to iden­tify who has made key fi­nan­cial de­ci­sions that go wrong.

“If AI and ma­chine learn­ing-based de­ci­sions cause losses to fi­nan­cial in­ter­me­di­aries across the fi­nan­cial sys­tem, there may be a lack of clar­ity around re­spon­si­bil­ity,” the re­port said.

Rev­o­lu­tion

The pace of tech­no­log­i­cal ad­vance will also make it harder to fash­ion durable rules for AI ac­tiv­ity that some aca­demics ex­pect to rev­o­lu­tion­ize the fi­nan­cial sec­tor.

The re­port said that RegTech in­vest­ment, or use of ma­chines to com­ply with a wel­ter of new reg­u­la­tions in­tro­duced to tackle money laun­der­ing and make banks safer, could reach $6.45 bil­lion by 2020.

Con­sul­tancy Ac­cen­ture said in May that three quar­ters of bankers sur­veyed be­lieved that AI will be­come the pri­mary way banks in­ter­act with cus­tomers within the next three years.

Euro­pean in­sur­ers in­vested $400 mil­lion in “In­surTech” or real-time tech­nol­ogy to help to re­duce pay­outs.

Nordea, the Nordic re­gion’s big­gest bank, said last month that au­to­ma­tion would help it to shed at least 4,000 staff. It has al­ready in­tro­duced an AI chat­box to an­swer cus­tomer ques­tions.

Dutch bank ING wants to in­crease the num­ber of traders us­ing AI.

Fund man­agers are also us­ing out­side spe­cial­ists to ob­tain ma­chine- learn­ing tools that sift through news and re­search for in­sight into mar­ket trends.

So-called quant funds use AI to man­age $1 tril­lion in as­sets. Though that is only a frac­tion of the $40 tril­lion in mu­tual funds glob­ally, the FSB said in­dus­try es­ti­mates sug­gest that could grow rapidly.

The reg­u­la­tors them­selves are also us­ing AI to make it eas­ier to de­tect fraud and money laun­der­ing, while cen­tral banks ex­pect to use AI for re­al­time pre­dic­tions us­ing big data to help de­ter­mine mone­tary pol­icy.

The FSB ac­knowl­edged that AI is help­ing the fi­nan­cial sec­tor to cut costs, im­prove prof­itabil­ity and widen choice for cus­tomers, but added that it also raises con­cerns over pri­vacy of data used in AI ap­pli­ca­tions.

Photo: IC

Aurora Chiquot (right), com­mu­ni­ca­tions man­ager at Alde­baran SoftBank Group, hugs the com­pany’s hu­manoid robot ‘Pep­per’ at the Alde­baran stand at the CeBIT com­puter and tech­nol­ogy fair in Hanover, Ger­many, in March 2016.

Newspapers in English

Newspapers from China

© PressReader. All rights reserved.