Tag Archives: robots

ROBOT JOURNALISTS THAT NEVER SLEEP

If you have daughters of a certain age then you too may have come across a stinking corruption of literature known as the Rainbow Magic Fairy Books.

Each lazily contrived volume is an attempt to come up with a new angle for the pretty standard fare of your common-or-garden fairy character, and it feels like meagre exaggeration to suppose that – positioned end to end – the entire compendium would stretch to hell and back. Having been begged by an excited 6 year old to pollute their absorbent mind by reading out loud one such example of this rotting crap, your heart and soul should be sufficiently burnt out to avoid exposing them to another. But you love your kid right? So that’s why I read another.

The unsurprising realisation that Kate the Royal Wedding Fairy’s story was as depressingly formulaic as Adele the Singing Coach Fairy caused me to wonder who on earth had written this nonsense. There are no authors credited on the spine, yet endless supplies of fairy mini-series pour forth; Sporty Fairies, Ocean Fairies and (I shit you not) Pet Keeper Fairies being just the tip of the iceberg. There was only one conclusion to my mind: they are written by robots.

Robot writers are becoming increasingly popular, as I read with interest in this chilling Guardian article by Yves Eudes this week. By 2025 about 90% of the news we read will be generated by computers, claims one of the foremost peddlers of automated writing technology. Well I hope to goodness I’ll be spending 100% of my reading time on the other 10%.

PR practitioners have grown used to the strange world of news aggregation websites that regurgitate enormous amounts of announcements from PR-land, despite being entirely bereft of human journalists. It used to be that these places just cut and pasted newswires to create a win-win situation for all involved: a win for the PR sender and their client because it’s an extra piece of exposure on the Internet for no cost or effort, and a win for the site owner whose traffic and corresponding advertising revenues would marginally increase. All of this was before Google changed the rules on its algorithm and now prefers ‘original’ content instead. Thankfully, the new Google regime supports the ideal of employing a human writer to interpret the more worthwhile PR blurb, apply their own news values to it and then publish the resulting original content as something someone might value reading. That content wouldn’t have existed if the press release hadn’t provided the impetus for the story.

I know a colleague who had a bit of a run-in with one of these automaton sites recently; let’s just call them storagegubbins.org. Rather than cut and paste, and rather than spend money on a writer, it looks as though they employ a rudimentary auto-thesaurus rule in a bid to fashion something different out of a bog standard press release; different enough to fool Google anyway. My colleague had put out their release about ‘ABC Components’, which read a bit like this:

“ABC Components, a major manufacturer of next generation chipsets, has announced a strategic partnership with BigBang Computers to extend their dominance of the European cloud IT market….”

Within minutes the website had published something more like:

“ABC Components, a chief industrialist of next generation chipsets, has proclaimed a deliberate syndicate with BigBang Computers to prolong their supremacy of the European cloud IT bazaar….”

I found this website doing something that looks alarmingly similar.

Back to the Guardian article, it’s hard to argue with the intelligent claims of experts but I can’t be the only one who thinks the idea of so much machine-driven writing is very sad indeed. Reading good writing is a joy, because you’re sharing the human experience. I also believe great writing is a performance in itself, where the words on the page conjure images in your imagination and generate some kind of emotional response.

I don’t know how machines can respect their readership. I just hope that it’s the sheer volume of data increasing to such as extent that only machines can supply the gaps in ‘news content’ that we don’t really care about anyway.