Devil's Advocate: Can we trust banks?

Mechanised decision-making, whether through computers or call-centre scripts, is good for many things - but not for interacting with human beings. Martin Brampton wonders whether businesses are eroding customers' trust by dealing with them in an automated manner.

Bank managers have never quite been the cosy figures of nostalgia. In the past they have been respected members of society, though, and projected an air of security and reliability.

Things have changed in recent years, and I wonder whether banks are now in danger of undermining the trust they once worked so hard to achieve. A reader of last week's article on chip-and-PIN, for example, noted that banks might very well attempt to make card holders liable for fraudulent transactions.

The banks' argument would be that a PIN is secure so long as it is not divulged. They made this argument for cash withdrawals over a long period, denying that it was possible for a criminal to fake both card and PIN. The denials went on long after it had been demonstrated that money was being withdrawn from accounts without the account holder's connivance.

In fact, the banks stand to get a good return on the investment in chip-and-PIN regardless of how much it may improve security. At present, most of the liability for fraud falls on the banks, with only a small proportion hitting the merchant. As the new system comes into effect, the liability is moved from the banks to the merchants - or perhaps back to the card holders, if blame can be pinned on them.

Maybe large organisations have always been economical with the truth in the way the banks were over ATM cards and 'phantom withdrawals'. Or is this part of a wider trend, encouraged by the ever increasing use of call centres? And could moving the call centres to the other side of the world be a sign of the detachment of front-line services from the core of the organisation?

Think for a moment what is involved in following rules. For people, it is a complex activity, and the borderline of what counts as following the rule is often uncertain. Take the case of Enron. Nobody denies that investors were deceived. It is much less clear who, if anyone, actually broke the rules governing company accounts - despite the 100,000 pages of explanation for the US version of the rules.

In some ways, this flexibility in what counts as following a rule has value. We would like to deal with organisations that can relate to our individual circumstances and judge accordingly. The quality of the judgement may be poor, or erratic, but we can feel that at least it involves reflection on how best to apply the rule on the particular occasion.

When we talk of computers making decisions, the situation is different. The process is mechanical, and given the same inputs, the decision will always be exactly the same. Call centres are constrained to operate very much like machines and not much like human decision makers.

In fact, the situation of a call centre operator is a curious contradiction. The work calls for emotional skills that are aimed at making the interaction successful in human terms. Yet the actual job of the operator is to apply strict rules without any genuine human input. And even where there are still bank branches, the staff are increasingly detached from management decisions, in much the same way as call centre operators.

What are the consequences of fronting an organisation with an apparently human face that actually lacks the real human decision-making faculty? It may make it easier to put out misleading information without the people delivering it feeling a sense of responsibility.

Maybe that would account for the experience I recounted in an earlier column, in which a whole series of customer services staff assured me that a major bank did not receive any financial benefit from operating 0870 numbers. Presumably this was the standard 'rule book' answer to questions about such numbers. Yet somewhere in the bank were the commercial people who negotiated the best deal for sharing in the revenue from those calls.

Or does the detachment of the front line from the inner workings of a bank make it easier to sell payment protection schemes on loans? They have been shown to be extremely profitable and have been sold to people who are very unlikely to gain any benefits.

Computing is the making of mechanical decisions at lightning speed. It can provide us with all kinds of benefits in its varied applications. But does it need to take us into a world of impersonal transactions that undermine trust?