On Saturday Prepare for a tsunami of virtual unreality

Read Martin’s column every week in the Western Daily Press

ATWO-MINUTE video about an old bloke walking through wintry British countryside might not sound like much, but watching it has caused my jaw to drop. Why? Because Artificial Intelligence robots made this little movie for me after I’d prompted them with a sentence or two on the theme and how it should look.

After my basic instructions, the AI app got to work sending me various messages about how it was finding the right kind of music and searching for relevant video clips and goodness knows what else. It then announced it was arranging a script and a voiceover.

Just 75 seconds later, there it was playing on my screen. An entire finished video.

I hadn’t moved from my chair. None of the footage came from my own hard discs – AI had “borrowed” relevant stuff from everywhere and anywhere on the internet including shots of some old guy walking in a lane. It had assumed the sort of mood I wanted and scripted a cliche-ridden voice-over, delivered in a rich male actor’s voice, backed by atmospheric music. Okay, so it wouldn’t win any awards, but it was, arguably, a better production than perhaps nine out of ten of us could put together. And this is the very first week that this particular AI app has been unleashed on the world of movie-making! The robots will become much, much better at it within months or even weeks. Which is scary, because it means that soon even an idiot will be able to “create” convincing videos about anything they choose.

That, in many ways, should be exciting. But it will take the question “can we believe what we see?” to a whole new level.

Part of the reason my jaw dropped was that, having watched the video, I could see there was a link between my little “old man in a lane” production and the Post Office debacle. I wrote about that in last weekend’s column and received a record breaking number of missives from readers – but I mention again because the entire fiasco centres around something which appeared on computer screens.

Post Office bosses obviously looked at Fujitsu’s Horizon records on their screens and concluded: “Yes, we believe what we see.”

They may have added: “It is a computer programme – computers don’t lie, but humans do.” Before getting their heavy-handed “investigators” to stitch up an innocent bunch of postmasters.

There have undoubtedly been other cases where dodgy computers have been believed above and beyond human beings – but after I wrote last week’s column about the emotive issues raised by the ITV drama, I began to see the whole disastrous shooting match for what it really was… by which I mean – not only a case of Mr Bates v the Post Office – but Humankind versus Computer.

Is it the first time a major nationwide computer glitch has occurred affecting the lives of thousands of innocent people? I don’t know. But it is certainly a glitch that shouted: “The computer says NO! You are NOT innocent. You took the money and you are going to jail!” It’s as if the powers-that-be were saying: “Computers are clean as a whistle because they only work with numbers and logic. Humans are endlessly complex and fallible – a few are greedy, conniving and utterly lacking in morals – something which a computer can never be. No wonder we believe the evidence we see on a screen!”

I have a close friend who is an experienced computer systems troubleshooter and his take on the Post Office scandal was: “Believe me, this stuff is happening all the time. It’s a miracle we’ve not seen other cockups on a similar scale.”

He told me he has often witnessed bosses hiding their tracks and covering their backs when a major computer glitch has hit the fan.

“These people are only interested in performance bonuses and the next lucrative job offer they can grab off LinkedIn,” is what he said. “They are quick to close ranks when there’s been a systems failure – it’s usually some poor schmuck down at grassroots level who gets the blame.”

All of which is unfortunate enough, but what are things going to be like if the work done by computers is enlarged a thousandfold and they are empowered on a vast scale?

True, they won’t be lured by performance bonuses or the promise of lucrative jobs, but they will make mistakes.

I know that because for the past year I have been fascinated by AI and experiment with it almost daily. Yes, it can be useful – but you have to fact-check everything because its rapid and seemingly learned pronouncements are riddled with weird and not-so-wonderful errors.

I dread to think what sort of highly convincing nonsense the new movie-making robots will be coming out with in a year or two’s time. We are entering a world in which we will not be able to believe anything we see on a screen.

The unfortunate postmasters were hit by some incorrect numbers, which was bad enough.

In the very near future, entire populations might be treated to authentic-looking news reports, documentaries, political broadcasts and so on which have absolutely no bearing in reality.

We need to be prepared for the tsunami of lies and inaccuracies which could be coming our way.