Banking

Computers are no Dickens but the writing's on the wall for human authors

We used to assess a piece of writing by the standing of its author and its source. But that will no longer be the case when a machine has done the wordsmithing.

Written in Tokyo and dispatched to TechRepublic a day later from the lounge at Narita Airport at 22Mbps via a free wi-fi service.

Computers will never play chess. They will never reason. They will never be creative. Those have all been commonly held views over the past 50 years.

Yet none of those assertions has stood the test of time, and our machines have surpassed us in many areas - including manufacturing and production control, games, general and specific knowledge, medical diagnosis and business modelling.

In the technology sector, computers have been geared up to solve problems, initiate designs and invent solutions. But in the financial and banking sector they are now seeking out data and compiling reports. And these are the reports that you and I get to read.

Can we tell the difference? Granted, financial data and forecasts are about as dry a subject as you can imagine, and the people who had been producing them didn't exactly have to be Mark Twain, Charles Dickens or William Wordsworth.

So, here we are with computer algorithms translating company reports and accounts and sports statistics into readable prose for ordinary people. Does it matter? I reckon not.

They will no doubt be more consistent, diligent and honest than their human forerunners, but on the downside they will also be capable of making unknown systematic mistakes that are frozen into their base code by human programmers.

As ever, there are tradeoffs between human and machine. People get tired, make mistakes, and do bad things, while machines slavishly battle on with unquestioning energy, breadth of field and depth of analysis. They never get tired, take a day off or make a mistake - within the bounds of their programming, that is.

This technology is in its infancy today but it is learning fast and the migration is likely to go something like this: finance, banking, insurance, sports reporting, road traffic, train and air transport, global and regional news, medical assessments and so on.

I really can't see any sector escaping this trend and I can only imagine the relief it will bring to those who currently face a blank screen and writer's block.

Personally, I have no philosophical objections - or ones of any other kind, come to that - and I suspect that it is only a matter of time before machines take a news story and turn it into a book, or a book into a movie script.

The clock is ticking as they say, and I just hope we get a veracity checker in there at the same time. Who knows: this might be the way to get honesty and fair-dealing back into a number of sectors - by pushing the people out and the machines in.

About

Peter Cochrane is an engineer, scientist, entrepreneur, futurist and consultant. He is the former CTO and head of research at BT, with a career in telecoms and IT spanning more than 40 years.

12 comments
peter
peter

You are correct...and time seems to be getting shorter...when I look at all the AI domains...the progress now being made is really dramatic. The change in the next 50 years will eclipse the previous 5000 in this field and many others....

homemade
homemade

Financial data and forecasts are intentionally wysiwyg. Just one level. (unless you count "Buy" or "Sell" as a second level) No sweat for a computer I suppose. Interesting novels do more than one thing. They may address more than one audience. Adults and children. (Dickens, Hard times) They operate at more than one level: An adventure story AND a call for more humane schooling (ibid). They can describe and comment on the human condition from the inside. Eventually it won't be beyond a computer to write a novel given some hints about what the deeper layers should be. Computers can already write poor jokes, this is a first step towards understanding layers of meaning. It could take a loong time to get a computer interested in deciding for itself what was interesting about human politics, (marriage and) social change or power, and a bit longer to be motivated into turning that interest into a novel.

Deadly Ernest
Deadly Ernest

now writes fiction, I can tell you that I see no need to worry about computers writing reasonable fiction for a LONG time unless there is a quantum leap forward in the development of artificial intelligence well beyond what anyone expects at the moment. Most financial reports are in a series of set formats with very basic layouts and content, so easy to do. I'd love to see if the current software is any good at doing the fictional style financial report needed for things like government agency estimates and additional estimates to get support for project funding, or doing the creative work sometimes needed to account for money that got blown because some big boss screwed up and we're not allowed to blame them in the report - mainly because they're the one it's going to. Anyway, all of those are very different to writing a good and entertaining fiction story.

VosaBz
VosaBz

Dear Peter, have you read Nassim Nicholas Taleb? Every analysis is based on data of the past. There are areas where predictions made on these data work quite well. But surely not in finance, and sociology, and many other areas. The future is uncomputable, otherwise our lives would be uninteresting. Have a nice day.

peter
peter

So we had better look out for the film script, the avatars and the computer produced movies. Many thanks for the input!

sysadmn
sysadmn

http://www.wired.com/gadgetlab/2012/04/can-an-algorithm-write-a-better-news-story-than-a-human-reporter/ " Every 30 seconds or so, the algorithmic bull pen of Narrative Science, a 30-person company occupying a large room on the fringes of the Chicago Loop, extrudes a story whose very byline is a question of philosophical inquiry. The computer-written product could be a pennant-waving second-half update of a Big Ten basketball contest, a sober preview of a corporate earnings statement, or a blithe summary of the presidential horse race drawn from Twitter posts. The articles run on the websites of respected publishers like Forbes, as well as other Internet media powers (many of which are keeping their identities private). Niche news services hire Narrative Science to write updates for their subscribers, be they sports fans, small-cap investors, or fast-food franchise owners."

peter
peter

When it comes to entertaining fiction I sometimes feel that a machine would be more imaginative, and certainly when I watch TV or see the output from Hollywood my suspicions in this regard are usually confirmed :-) It will be interesting to see where this goes in the next 50 years !

peter
peter

But that's what bankers, investors, pension funds and insurance companies do. That is the basis of the finance industry. Rightly or wrongly - that is what happens - that is what is done! I never suggested it was right, proper, correct, or accurate :-)

peter
peter

Thanks for picking up on this article in Wired. I seems to say it all - and as our education systems deteriorate our machine will look even better. Those excerpts look pretty good!

Deadly Ernest
Deadly Ernest

already being semi computer generated via pre-set formats and formulas of what some idiots in those industries think will sell. Sadly, due to the lack of good material it sells simply due to there being nothing else available. Most financial reports I've seen have been of the style of 'this is what's happened over the last period' and been extremely boring. What does get interesting, especially in government circles, is the creative writing to justify extra funds for new small projects and replacement equipment. Now they can get real interesting reading in seeing what people come up with, and I ALWAYS had fun writing some of them, but they were few and far between. For the best entertainment reading now days you have to forget the best seller lists (amazing the amount of rubbish on them due to formula style writing) and look across the web for amateur and semi-amateur stories from places like Fine Stories, Stories On Line, Lulu, Dpdotcom, and other e-publishers and self publishers. One word of warning, the Kindle, Nook, e-pub, and similar e-reader formats are crap go for html or PDF as they have reasonable formatting and can be easily read.

peter
peter

Why am I not surprised ? And thanks for the input and tips!