Séamas O'Reilly: If it feels like scams are thick on the ground - it's because they are

Seamas O'Reilly. Picture: Orfhlaith Whelan

This week, a man in Hong Kong set a brave new benchmark for bad days at the office.

At his job in a finance company, he was called into a Zoom chat with his colleagues, including the chief financial officer of the entire firm.

In this meeting, they laid out a plan to transfer a very large sum of money to five Hong Kong banks across 15 secret transactions.

The employee says he found this request alarming, but dispelled any doubts he had once he logged into the conference call and spoke with colleagues he knew and recognised, as well as the chief financial officer, whose stature and seniority far outstripped his own.

His orders given, the employee made the transactions and heard nothing more about it until he made inquiries with the company’s headquarters a week later. 

It was then that the extent of what had happened became clear. 

Not only had the company not ordered these transactions; they had never even held the meeting. 

The entire thing had been a ruse and every other person he had seen and spoken to on the video call had been a deepfake impersonation. 

He had, in fact, just sent 200m Hong Kong dollars (€23.8m) of the company’s money to a squad of online scammers.

Anyone reading that story might feel a slight tightening in their chest. 

If you’re anything like me, you might also feel an amoral, but completely undeniable, sense of amazement at the scammers’ audacity and ingenuity — the palpable thrill of a bank heist at its most futuristic. 

Any such sense of schadenfreude must, however, be tempered by something else: a sinking sense that this is merely the tip of the rotting iceberg of scammery we all now inhabit.

It feels like scams are thick on the ground because they are.

Almost no one escapes interaction with them in 2024. 

When I began writing this piece, I realised I hadn’t had a spam call in a while, and promptly received one five minutes before I sat down to write this very sentence. 

It was of the standard variety; a robotic voice that was about to tell me to seek compensation for an accident at work before I hung up. 

We all know people who get these daily. My dad gets them constantly, as I discover any time I call him and receive a gruff ‘hello!’ because he’s presumed that I’m the tenth such call that afternoon.

My mother-in-law routinely gets texts from ‘An Post’ telling her to pay for packages from America, where her sister happens to live and often dispatches gifts. 

Numerous banking, crypto, insurance and utilities scams operate by meticulously spoofing official websites, emails, texts, and customer service lines.

Telecoms regulator ComReg reports that such fraud is thought to cost people €300m a year in Ireland alone, and this is likely a low estimate. 

Online, of course, this situation is simply more accelerated. 

Sometimes, the effects are merely comical. Last week’s edition of this column was an explainer on the Stormont deadlock, on the eve of its being broken by the DUP. 

Now, I will admit I considered this a perfectly lucid and entertaining bit of writing, but an X user was so moved by its analysis that she replied, immediately, with a payment link to videos of her porn.

Such tactics are not new, of course, but they are much more regular than they were even a year ago, having been granted greater license by slack regulation. 

To take the above incident as just one example, that scam bot had a blue check mark, meaning that, unlike me, it pays money every month to Elon Musk’s vastly indebted and unprofitable platform, a situation which would greatly disincentive his company taking proactive measures to weed them out. 

This reluctance might also be observed in the proliferation of garbage in Google’s search results or Amazon listings.

Secondly, tools like OpenAI can create bots for free in seconds, in their thousands, as well as things for them to say, and pay links for them to share. 

This has created what has become known as The Dead Internet; huge swathes of the web comprised entirely of bots, often talking to each other, in a never-ending churn of contentless garbage which may now outnumber the population of actual humans.

Sometimes, the mask slips. If, for example, the language modeller OpenAI is incapable of fulfilling a task, it returns the message ‘sorry, I cannot generate the requested content as it violates OpenAI’s use case policy’. 

Doing a search for that term returns tens of thousands of accounts, posts, products, reviews, listings and personal websites — all fake, and growing in number every day. 

And this, you must remember, is just a tiny percentage of the full total, glimpsed solely through the window of just one infrequent error in just one of several AI platforms. 

Since there is no meaningful upper limit to the creation of these bots, the scale is likely too staggering for us to comprehend. 

To put it another way: if creating a scam bot has only a one-in-a-million chance of making you money, that’s merely an incentive to create a million of them.

This is the problem with a phenomenon as vast and multifaceted as scamming. 

While we’re rightly terrified and fascinated by stories about sophisticated deepfakes robbing Hong Kong banks of millions, a more mundane form of the AI apocalypse is already here for the rest of us, and has been for some time.

The future’s true footprint is in the deadening grind of bots now littering every single layer of the internet, talking to us, talking to each other, on the off chance they might one day, somehow, trick somebody, anybody, into giving them a few quid. 

It’s selling you a fake version of my book on Amazon, it’s replying to your social media posts with porn, it’s texting your mam about her packages, it’s calling up your dad about his car insurance and it’s going nowhere. 

The future is here, unfortunately, and it’s a lot more boring than a deepfake bank heist, but no less terrifying.