You are here

Oh Brother, Big Data

It’s 2017. Corporations say they’re using data analytics to improve society. Or is it 1984?

BT_20170415_COVER15_2840760.jpg

BT_20170415_JLDATA_BY_2840854.jpg

BT_20170415_JLDATA_ICON2_2840792.jpg
"You have safe, and you have 'creepy', where it's easy to say 'ok' or 'cannot'. And the truth of what we have to deal with lies with the nuances in the middle." - Lam Chee Kin, group head of compliance at DBS

BT_20170415_JLDATAUIO1_2839508.jpg
Grab's data also yields surprises, such as the fact that many part-time drivers are not chasing incentives. "They'll say, 'I'm doing this for fun. My wife wants me out of the house'," quips Lye Kong-Wei, head of data science at Grab.

BT_20170415_JLDATAUIO1_2839508.jpg
"Nowadays, what the KGB would have gotten out from torture, we reveal it on Facebook. Have you tried to live without Google for a week? So, embrace it." - Andreas Weigend, director, Social Data Lab

WALLS are popular these days, and Andreas Weigend knows all about them. In 1949, the East German's father was thrown into jail by the government on suspicions of being an American spy - likely because he spoke English. After the Berlin Wall collapsed, Mr Weigend asked for a copy of his father's file. The Stasi disappointed the former chief scientist at Amazon. "'We're sorry, we've destroyed the file on your dad. But since you asked, would you like a copy of your own file?'," he recounted at a Big Data conference in Singapore in late March. As it turned out, the Stasi had been gathering intelligence on him, too.

Mr Weigend advocates the use of Big Data, which may seem odd considering that an overbearing watchman had wrenched his father from his life. Yet, as Mr Weigend lives through Germany's reunification, and surveys a world transformed into one with low barriers of entry to information from most corners of the world, he says it is inevitable. Walls must come down.

"Nowadays, what the KGB would have gotten out from torture, we reveal it on Facebook," he deadpanned at the conference held by Agoda.

"Have you tried to live without Google for a week? So, embrace it."

sentifi.com

Market voices on:

But as businesses and governments begin to tap on Big Data and sidle up to consumers and citizens, there is a growing need for a new social contract.

In the area of trust, technology companies appear to have won the battle. Over the last six years, trust in technology firms has consistently stood around 75 per cent, the Edelman Trust Barometer this year showed.

Google now knows where you work and live, what piques your curiosity, and whom you email frequently. Amid that broad trust, many consumers today simply do not read the terms and conditions attached to the use of data by social media and technology companies, observes Lam Chee Kin, group head of compliance at DBS. "Social media platforms and tech companies extract very wide consents for use of data."

Edelman also noted that other industries are gaining more trust from consumers than before, while trust in technology companies has plateaued. "People are not convinced that the tech sector is adequately transparent and authentic in how it operates, nor are they clear on how technology is contributing to the greater good," says Michael Stewart, global vice-chairman of Edelman, in a report this year.

Economics 101

But technology firms insist they are doing work for social good. For rail-hailing business Grab, a zero-sum analysis of its business is limiting - its business is to boost the efficiency of cars, says Lye Kong-Wei, head of data science at Grab. "The general opinion is that we're trying to steal demand from public transport," he says.

"We're not out to gouge anyone, we want to be fair to passengers and drivers. Dynamic pricing is a function of imbalance of supply and demand. It's Economics 101."

To clear the market, Grab gathers data for analysis. And it says it has been pretty accurate "on good days" in predicting and directing drivers to areas with high demand, without citing figures. With a network of more than 50,000 drivers in Singapore alone, it has tried to meet drivers' needs as well, gauging the odds that a driver would bid for a job based on historical and real-time data.

Besides getting more accurate matching, Grab has also cut down the number of unwanted jobs that appear on the drivers' app, after drivers flagged safety issues from seeing multiple pings on their phones. In Indonesia, it notifies bike riders when they are speeding, leading to a 35 per cent fall in speeding incidents.

Its data also yields surprises, such as the fact that many part-time drivers are not chasing incentives. "They'll say, 'I'm doing this for fun. My wife wants me out of the house'," quips Dr Lye.

Grab runs experiments, which is why two people going from the same point A to point B may be offered different prices or rebates. About half of the data work is through experiments, with the rest being simulations. Grab is clear that experiments are not only expensive but also "mess around with people's relationships", and says it only uses experiments when it has no choice. It says it doesn't keep personal data "in perpetuity".

But as Grab doubles its staff size from a year ago, it may eventually have to come to terms with the rising criticism faced by its global compatriots. Google is under fire after online ads by big brands were found to have been placed alongside extremist YouTube content that included racist and homophobic slurs.

In the wake of backlash over fake news on social media, Facebook chief Mark Zuckerberg penned a 5,700-word manifesto, pledging to build a global community amid fears that ideological lines are being hardened and shaped by algorithms.

The manifesto has drawn scepticism. "It will take more than sharing funny cat videos to create an effective global revolution," the historian Yuval Noah Harari wrote in the Financial Times. "You can hardly lead a global community when you make your money from capturing people's attention and selling it to advertisers."

Cool or creepy?

Banks may find some relief as the industry's trust score - according to Edelman - has been trending higher, even as the financial sector remains far behind the technology industry in winning trust, following the financial crisis. With consumers wary of their predatory ways, banks are aware of the goldmine that data is, but are cautious around it.

More than one bank operating in Singapore has a skull symbol flashing on desktops, warning bankers to be careful with banking data. Many bank employees still cannot access their work files from home.

Singapore lenders are bolstered by the top marks that the country scores as a financial centre, according to Edelman. But the swirl of complexity behind engaging banking clients means that banks are still wary of the risks of being too close to its clients.

"Our privacy promise to our customers builds on the idea of trust and responsibility. We have to start from that: the use of data as a stewardship of our relationship with our customer. It sets an ethical or moral foundation for what you're trying to do and tries to express some trade-offs," says DBS's Mr Lam.

DBS has launched several Big Data experiments which have tidied up its ATM network, and automated some of its branch auditing. But Mr Lam says banks are limited by their need to protect their reputation and by their duty to society.

"I could intentionally hire ex-ACS students as bankers specifically to target prospects or customers who are ex-ACS in particular professions, segments of society or particular addresses, and set them modified sales targets accordingly," he says, while making clear DBS does not do this. "You have safe, and you have 'creepy', where it's easy to say 'ok' or 'cannot'. And the truth of what we have to deal with lies with the nuances in the middle."

OCBC used to send messages to customers the second it sees where they have swiped an OCBC credit card. But the bank has scaled back on such rapid location-based marketing, because it can get "a bit creepy", says Donald MacDonald, OCBC's head of group customer analytics and decisioning. "I can see that you've had a very large deposit yesterday. But when a staff calls, he doesn't say, 'Hey, I see you've put in fifty thousand dollars'. We'll be a bit more subtle."

But it's hard for banks to keep their hands off data analytics. About a decade ago, OCBC was using hypothesis testing for its marketing campaigns, and the successful sales conversion rate was a mere 1.5 per cent. For some of the bank's real-time campaigns today, conversion rates have gone to as high as 60 per cent, with the average at about 22 per cent, says Mr MacDonald.

Banks are also monitoring their staff more closely, using Big Data to track retention figures, and for staff welfare. OCBC has looked through credit records of staff to offer emergency loans. "So if you're having financial difficulties, don't go to a moneylender down the road," says Mr MacDonald.

But banks are likewise using Big Data to prevent internal fraud in order to keep their reputation untainted by bad bankers' behaviour in a post-crisis era. "In the case of banks which have a public interest to serve... there arguably exist even stronger reasons for the banks to monitor the flow of information and correspondence for questionable activities," says Ian Lim, TSMP Law's employment and labour practice head. Just last month, an ex-investment banker was fined in the UK for using WhatsApp to leak client information.

Anonymous is as anonymous does

In Singapore, businesses are expected to destroy personal data collected once there is no longer a purpose for its use. If a person has gone for a job interview and provided information for it, the purpose ceases to exist when the interview is done, and he does not get a call back. That data should not be used to extrapolate the types of candidates who fail at interviews, for example.

That argument should still hold even if companies try to use the data by "anonymising" it. Tan Shong Ye, digital trust co-leader at PwC Singapore, points out that Big Data, in the hands of an organisation with enough scale, is still personal data that can be linked to a single identity. "Can Big Data be fully anonymised? It can even be argued that once you've collected enough data, then every person is unique."

And while businesses should be adopting win-win approaches over the long term, they are prone to short-term pressures. "If improving lives means good business, then the two are not necessarily in conflict," says Tan Ern Ser, associate professor, department of sociology, at the National University of Singapore. "Being a good corporate citizen is good PR. However, I believe profits usually trump other considerations."

Dr Harari worries about the "neoliberal cocoon" of technology companies. But that cocoon seems to be unravelling. Facebook has added a new feature in the US, where users can follow and contact government representatives.

Grab, which is backed by Temasek Holdings, is aware that regulatory support rests on the company keeping to its socially responsible tune. The Singapore-based company has sent transport data to governments, and participated in a World Bank study on transport systems in South-east Asia. It is also collaborating with GovTech over a shuttle bus service.

The creeping fear is that the Big Data cooperation between businesses and government has a tinge of moral policing. Alibaba is working on a credit scoring system where consumers could be rewarded for good behaviour on social media. Mr Weigend is looking at Alibaba's Sesame Credit rating with interest.

But he also notes that the US government has subpoenaed data from Amazon's voice-recognition assistant, Alexa, over a murder case - Amazon had relented when the defendant agreed to hand over the data. Mr Weigend is able to contemplate a day when the government could get a recording of his bedroom activity.

Yaron Zeidman, chief technology officer of Agoda, is more blasé about Orwellian surveillance. "It's not like the government hasn't been listening to our phone calls in the last 30 years. When they really want to get some data, they can get it."

PwC's Mr Tan notes that Singapore has an emphasis on data protection, as opposed to privacy. "It doesn't spell it out as an individual right unlike in the West," he says. The Personal Data Protection Act (PDPA) only specifies what can and cannot be done with the information, and when informed consent is required, adds TSMP's Mr Lim.

The Personal Data Protection Commission told The Business Times that it is looking into the consent regime. It declined an interview on Big Data, saying the review is "in its infancy".

"The question of actual ownership of personal data or Big Data is a complex and involved issue at law that has not been fully answered to date," says Mr Lim. Furthermore, PwC's Mr Tan notes, the PDPA excludes the government.

In 1651, English philosopher Thomas Hobbes published his classic on social contract theory, where he famously described life as "solitary, poor, nasty, brutish, and short", arguing in favour of strong, absolute, leadership.

As Big Data forces society to confront its own nasty and brutish ways, a new Leviathan is emerging. It is built from a montage of governments and businesses that smile and say that they are here for your own good. Any choice in the matter looks increasingly limited. Progress is presumed.

Powered by GET.comGetCom