In the 1980s, I was sitting in the office of a software company. They were an early supplier of legaltech and I was an early customer. We were, figuratively, shredding the software their data and systems people had built and rebuilding it as a combination of how a lawyer's office actually worked, rather than how systems people thought it should work, and developing the then novel idea of an expert system for litigation and conveyancing case management and support at a time that law offices used computers for accounting, time recording and word processing. Their system was as visionary as it was flawed. I wanted it fixed.

The head of tech support for the software company was in the meeting and when a frantic clerk pushed her head around the door about someone who was in distress, he had it put through and said "you need to know the kind of thing my department has to deal with so we can try to prevent it at the design stage."

The conference phone speaker was filled with shrieking; "My boss has gone mad. He's threatening to throw the computer through the window. Now he's punching the screen shouting "Do as I fucking tell you."

It was a toss up: did we tell him to send the computer back as he was obviously unsuited to have one, did we try to get him to talk to support and find out exactly what he couldn't work or did we send for the men in white coats?

We, almost all from the British Midlands, were blessed with a sense of humour that allowed us to make fun of things while trying to find a solution and to help the chap, not because he was a customer, per se, but because no one should be driven to distraction by technology. Sadly, today, forty years on, that is a lesson that tech companies have still not learned. 

Why? It's simple: computers do not care about you. 

You can't reason with a computer. 

"Psychopathy is a condition characterized by the absence of empathy and the blunting of other affective states. Callousness, detachment, and a lack of empathy enable psychopaths to be highly manipulative." - Psychology Today

In the past week, I have published two posts that demonstrate aspects of this:

https://www.linkedin.com/posts/nigelmorriscotterill_morriscotterill-freespeech-languagepreservation-activity-7324976018312585216-d_2L?utm_source=share&utm_medium=member_desktop&rcm=ACoAAAAM9AoBaYQ7sT56O-hlrCFYsXOYWWyKEbg

That's about the dangers of homogenisation of language, and the use of language, in a global, culturally diverse - diverse in every sense imaginable - world, so that the making of errors of judgement are enshrined in computerised routines. 

This was the other post:

https://www.linkedin.com/posts/nigelmorriscotterill_content-moderation-simulation-trust-safety-activity-7323993103755829249-5nN1?utm_source=share&utm_medium=member_desktop&rcm=ACoAAAAM9AoBaYQ7sT56O-hlrCFYsXOYWWyKEbg

This relates to the views of a small group of people as to the acceptability of certain language and how their own interpretations and ideologies create, not ease, division and discrimination. 

I took particular exception to the lack of grey areas: even mild insults are termed "hate". The consequences of being described as using "hate speech" can be catastrophic. Being kicked off LinkedIn, etc. for using English instead of American is bad enough but where those decisions are made by algorithmic processes programmed by people who design in their own prejudices, which often they minimise by referring to them as a "bias" the consequences become hazardous.

In their own use of language, they design out clarity and design in fuzziness but in their algorithms, they design out any form of empathy - as they must because computers cannot feel, they cannot be empathetic, they cannot respond to the sad faces or tears or fists of someone who has reached the end of their tether. 

This is where the use of computers is reaching a crossroads.

Some computer systems have been programmed to identify stress in the voices of those making false insurance claims, for example, and to identify tones of voice that identify excessive excitement by dealers who may be giving or receiving insider information. This, I would argue, isn't empathy but it is identifying factors that empathy identifies. But let's be absolutely clear: this is not in the vast majority of software that you and I interact with. 

And it's not in the vast majority of software that is making decisions that affect lives and livelihoods. Why? Because computers don't care because computers can't care. 

The computer doesn't feel pain when a frustrated user punches it. It might break but it doesn't take the afternoon off to visit the doctor or to make a complaint to HR. It doesn't even take paracetamol. Whatever you do, it makes no difference to the computer unless it has been told to respond in a particular way to something you do and that something is limited to defined inputs. 

It is important that we all understand this because increasingly our lives are managed by computerised processes which means someone, somewhere, has defined those processes and listed expected inputs. 

"Do as I F***ing tell you" - computer generated by and (c) vortexcentrum.com

Oddly, none of those inputs seems to be "do as I effing well tell you."

It's a good idea to visit the website of one of the companies that lets you set up your own chatbot, with a free trial. It's very challenging because if you are a human that does think about other people, then you will try to anticipate human responses. And then two things will happen. The first is you will realise that there is no obvious limit to what those responses might be. The second is that you will realise that for every way your system reacts to each of those responses, there will be a similarly unlimited range of responses to your reaction. Eventually, you will reach such a state of confusion and complexity that the only way out is "would you like to speak to a human?"

The tech world thinks it's moved on from automated phone answering systems but in truth it hasn't and we are all to blame for that illusion.

In the 1980s, offices generally had telephonists. But it wasn't long before internal phone boxes provided for direct lines without the need for the installation of additional wiring - until then a direct line was literally that: a wire from the exchange via the box in the street to someone's desk. Then came mobile phones and in many cases ego said to give the number to everyone. 

That was coupled with a change in phone use: previously, telephone calls had been expensive. But with cheaper calls, and direct calling, the idea that everyone was always available became ingrained. Callers became indignant if they were not answered. And with that we got individual answering machines that some wag termed "voicemail." And instead of saying "enough" we accepted it. 

But the voicemail systems, developed from "if you know the person's extension please enter it now" was the genesis of the telephone menu systems "to speak to Department X, press three." 

We are still in the same process: albeit bigger and more unwieldy. From trying to report a lost or stolen credit card to trying to find out how to pay an electricity bill, we are fed into tunnels of predetermined questions with limited opportunity to answer and no chance to break out when it's clear the questions are taking us down a blind alley. 

When we interact with a computer, our communication with it is constrained not by our own inability to communicate but by the imagination of the programmer or, just as likely, his decision that the chain of questions and answers is long enough and in any case it's time to go to the pub.

In the 1970s, in the British car industry, there was a recognised quality control problem: "Friday cars" would need to be screwed together properly by dealers before they could be handed over to customers. When companies tried to impose quality control measures on production lines, there was open rebellion and strikes - the 1970s equivalent of today's moaning about "toxic bosses." 

Toxic Technology

Today, as customers are constantly faced with technology that is not fit for purpose, when complaints are met with accusations, when people lose control over the most basic of human condition - that of dignified communication with a business that they spend their money with - in circumstances where rigid following of internal processes makes the customer's life unreasonably difficult; when banks, for example, have multiple systems to do the same thing and so what works today on one system will not work tomorrow on another, it's not at all surprising that customers, recognising that the computer doesn't care, want to take it out on the company. 

That might mean a whinge on yelp or it might mean a brick through a window. Or a vengeful hack. 

Computers don't care. Businesses must reverse that and if that means less automation, then so be it. But it also means less strict protocols adopted by humans. Discretion and common sense must be restored in the interactions with customers and others otherwise society is going down a very dark path in which, ultimately, ordinary people reach the point where their mental state is affected by the endless difficulty in trying to get the simplest things done. 

Nigel Morris-Cotterill is a financial crime risk strategist and has long identified technology as a significant risk factor not because people use it to commit crime (which of course they do but that's too obvious to bother thinking about) but because failures in tech undermine the relationships which discourage people from committing crime. 

www.countermoneylaundering.com