AI horror is already here!

Artificial intelligence, it’s amongst us, whether we choose it to be or not, it’s in our homes, on our phones and in our daily lives. But in films AI is less Siri? And more serial killer.

From murderous robots to big brother style overlords, Horror loves to imagine what could go wrong with this type of technology, sparking a new or potentially an existing fear into all of us. But how much of this is just popcorn fiction and how much could actually happen? 

The Origin Story of Many Horror Fans

I took my first stab at horror at the age of 8, a friend and I thought we were clever to sneak one of her mums VHS tapes, the chosen contraband was a nightmare on elm street, needless to say it was my first sleepover where I didn't sleep a wink.

Instead of scaring me for a lifetime, Freddy awoke a lifelong obsession with the genre, including its AI tropes, So naturally when I was offered a job at a tech company I was a little worried that my screams would now be because of the YouTube algorithm but the good thing about working in a tech start up... I'm surrounded by nerds who can calm my fears! I've actually asked one to help me with this video. Let’s head back to the office to meet him

One, Two, Freddy’s coming for you…..or is it Alexa?

One, Two, Freddy’s coming for you…..or is it Alexa?

So, Graham here, has agreed to help me put five typical AI horror tropes to the test, are they terrifyingly possible? or do they stay in the realm of science fiction? Shall we start?

I have a wheel thats going to help us do that, we have the five horror tropes on here, Killer Robot, Big brother, The Puppet Master, Deep Fakes and Cursed Technology. I will give an example of a film that represents that specific trope and you, Graham will give your opinion on how that could potentially happen in real life.

Lets take it for a spin…

The wheel of…. not so fortunate

The Killer Robot

The Horror:

Okay first trope on the list ‘The Killer Robot’….

Take the film MEGAN for instance, on the surface, it’s just about a cute little AI doll designed to help out a struggling family.

She’s meant to be part babysitter, part bestie and part bodyguard, the perfect solution for busy parents. But of course, it all goes horribly wrong when MEGAN decides she's not just protective, she's overly protective. And by ‘overly’ I mean breaking necks if you so much as looked at her kid the wrong way.

It’s the classic ‘technology gone rogue’ storyline, we invent something to make life easier and ironically it becomes more chaotic having to fight off that very thing.

And lets be honest who isnt making sure we say please and Thank you to our AI friends just in case they remember? I know i do, so how likely is this to happen in the future?

Just your innocent AI Nanny…. Right?

Graham’s Opinion:

Before the recent AI boom I would have said low, but now we’ve seen AI companies are happy to build intelligences as huge black boxes that they freely admit they have little understanding of. We wanted Asimov’s three laws of robotics, but so far all we’ve got is feeding a prompt asking nicely for them to do what we want, and we have ample evidence they follow these instructions … inconsistently. So I’m at a 7 on this one, pretty plausible.

Not sure about hacking machines wirelessly as you walk past them, though. And why do the creators of the hyper-intelligent robots always decide to give them super strength? We can definitely learn something from that Sci Fi trope!

Remember your p’s and q’s

I'm definitely lacking in upper body strength thats for sure. Let’s take another spin….

Big Brother

The horror:

So that was an example of a human intervening and making things worse, but what about the risk of an AI going rogue itself? We’ve all seen ChatGPT hallucinate.

The things I’m thinking of are the big brother in the cloud, watching and recording everything we say or do and using this data against us, blackmailing or exposing those deep dark webbed secrets to the point of no return, take the film ‘Unfriended’ for example that does exactly this!.

The whole film plays out on a computer screen, with a group of friends chatting over Skype when things start getting glitchy, suddenly there's a mysterious user logged into the call and nobody knows who it is. As the night goes on, secrets get exposed and tensions rise and one by one something happens to the friends on the call.

Could AI hack into our calls and start listing off all our secrets?

Is AI always watching?

Graham’s Opinion

That one is probably already happening, given humans have been performing these scams for years.

Scammers are early adopters of AI, and ones who don’t care at all about safety, regulation or international law. Right now they are busy using AI to scale up and automate their existing grifts, including impersonating authorities to defraud, blackmail, shame scams and so on. I think the general public are unprepared for the onslaught.

Less realistic is that AI will be able to gain access to data such as CCTV recordings that are private. The same protections we already use against humans should also slow down AIs to a similar extent, at least on current technology.

what can we trust?

Give the Wheel another Spin….

The Puppet Master

The Horror

So let’s make it even freakier then, what if we didn't limit AI to objects but gave it control over our own bodies? The Wheel has landed on ‘The Puppet Master’.

The film I'm referring to here is Upgrade, where a quadriplegic has a microchip inserted that not only restores movement but turns him into a full on fighting machine, with precision, speed and some terrifying perks, sounds like your perfect G I joe right? until the chip starts making decisions for him and suddenly he's not using technology, technology is using him.

And you have to wonder with gadgets such as fitbits and smart watches are we just an ‘upgrade’ away and could AI really be a master of mind control?

Think before you Upgrade

Graham’s Opinion

It’s maybe one of the less plausible ones, but definitely possible.

We’re right at the start of a world of Deus Ex style augmented humans, with the initial uses being about medical therapy, but with more everyday uses coming next, and inevitably military applications.

Current technology allows us to connect maybe 100 nerves independently between flesh and machine, and so we are getting close to the level where some hidden communication is starting to become possible. So I would say thats a credible threat.

On the other hand I have a fitness watch, and I’m not worried about that taking control anytime soon.

Can AI play us like we play the sims?

Phew it had put me off wearing it for a while.

Another spin on the wheel…..Okay let’s bring it back closer to real life.

Deepfakes

The Horror

They are here, we’ve all seen an advertising of a sequel that hasn't even been confirmed by the studio.

Or AI using someone's voice to create a disturbing phone call to a loved one in the hope of scamming someone. In fact, I got a text like that just last week. So like the film ‘Afraid’ could AI create fake videos, documents or posts that could have some serious consequences?

In the movie, what starts out as a friendly AI assistant quickly turns dark, using deepfakes to spread lies and completely mess with the characters sense of reality.

We are already in a world where we have to question what's real and what's fabricated aren't we?

Graham’s Opinion

I think this is painfully on point.

We have a constant war of attrition between users, who find security a frustrating distraction from whatever they’re trying to do. Meanwhilst, scammers, hackers and some governments use the international nature of the internet to evade consequences.

That is already ongoing, so highly credible, and the depth of these scams is going to explode soon.

Is that really the poster for that film?

Possessed Technology

The Horror

I’ve got one more category for you, possessed technology, like things that take away your agency or have been thrust upon you.

Take these three films for example all of them use this trope but over three different platforms: The Ring, a killer video tape that gets shown around or sent to people on the dark web to download and watch and seven days later your unalived, you’ve never wanted a working week to go so slow.

Then you have Stay Alive - a video game where the term ‘game over’ has a whole new meaning if you die in the game you die in real life and finally The Countdown, an app that quite literally tells you by the exact minute when you will meet your end, you’ll never trust a timer again.

But the likelihood of technology being taken over by something demonic surely that's out of the question?

7 days!….

Graham’s Opinion

Well, thinking about it, demonic possession is hardly a new risk.

Stretching a little, but it is possible AIs could hypnotise and plant suggestions in people through TV and video. Imagine an omnipotent Derren Brown which never sleeps behind every video screen! DerrenAI could do some serious damage.

And what about ‘The Countdown’, where you know exactly when you will die?

Finally, one I can debunk. Imagine manually setting the hands of your clock to the time you will die. That small movement of the hand will have sent millions of air molecules off on slightly different paths. That will grow through a butterfly effect to change the weather, and therefore many major and minor things about the world would change, inevitably including your exact time of death.

To fix it you need to set the hands of the clock to the new time of death, this will send the air molecules off in slightly different new directions …

Well that's reassuring.

The timer you wish you could hit snooze on

So, here's the bottom line.

Most of the scenarios we've discussed, They're not science fiction. They're not far-fetched. In fact, some of them are... inevitable.

And the root cause? We don’t have any real way to install “laws of robotics” into generative AI today. Sure, we can ask it to behave. We can suggest it follows guidelines. But let’s be honest we don’t know if it’s genuinely cooperating or just humouring us whilst working toward some unintended outcome.

The truth is, GenAI is still a black box. We know what we put in.  We can see what comes out. But what’s happening in between? We’ve imposed very little structure. It’s like a lab-grown brain, complex, unpredictable, and with no coded moral compass.

And that’s where we are. So where do we go from here?

Maybe we accept that Gen AI isn’t the alien, it's the mirror, we wanted it to think like us and now it does, art imitates life after all and if that's messy well it's only human.

Come on the journey with us by joining the Evil Lair.