Could our gadgets be listening in on us? It's not as crazy as it sounds...

"I have a crazy conspiracy theory," began a recent post on Reddit, "and I'm pretty sure I proved myself correct." Earlier that day, he or she had been chatting with friends in their living room about a particular club. Two hours later, an advert for that club appeared alongside their Facebook timeline. This, they thought, was spooky. They tested how spooky it was by reciting a list of words into their smartphone's microphone, including "Maserati" and "African safari". "Sure enough," they wrote, "two hours later I was getting ads for Gold Coast Maserati and African vacations… Do you think I'm nuts?"

Some commenters thought that yes, the most probable explanation was being "nuts". There's a cognitive bias, colloquially known as the Baader-Meinhof phenomenon, where we feel unsettled by something that appears with improbable haste after the last time we thought of it.

Humans underestimate the probability of coincidence, and as we don't know how rigorous the Redditor's experiment was, it would be hard to use it as evidence of a technological conspiracy. But plenty of people commented to the effect that they wouldn't be surprised – and that reaction is equally interesting, from a psychological perspective. There's now a commonly-held supposition, verging on assumption, that gadgets are listening to us. And worse, that they have nefarious motives.

Last week, a furore blew up around the terms and conditions of Samsung's Smart TV sets. "Please be aware," read one line, "that if your spoken words include personal or other sensitive information, that information will be among the data captured and transmitted to a third party." These TVs have a voice recognition feature which kicks in if you press a remote control button. Simple commands, such as "volume up", can be handled by the TV; more complex ones are outsourced to a server operated by a company called Nuance. As Samsung has no idea what personal secrets you might divulge during the few seconds when the feature is activated, it covers itself with a line in the T&Cs. But it was reported as if it were an Orwellian nightmare unfolding.

Nota bene: as it stands, in 2015, voice recognition requires the oomph of an external server to do the job properly. You could say that by calling their TVs "Smart" and burying the line about third party processing deep in the small print, Samsung is perpetuating a myth that the TV is doing all the work, and audio information isn't being sent anywhere. But that's how our new, lazy, voice-controlled world is facilitated, whether it's Apple's Siri, Google Now, Amazon's Echo or Microsoft's Cortana; a recording of your voice will end up somewhere. Someone might even listen to that recording; another post on Reddit this week came from someone whose job is to listen to clips from voice recognition services, compare them to text translations and assess them for accuracy.

Gadgets have to make voice recognition feel like magic, but it's just data being crunched in return for cash. We want magic; we don't want to know how it happens. When we're made aware of how it happens, we're forced to think about the journey that data is taking. From those thoughts emerge paranoia – and who's to say it's not justified, because the paper trail is far from transparent.

So we stare at the microphone, a pinhole recess in a piece of black plastic, and wonder if the advice always given to politicians – "assume that the microphone is always live" – might now apply to us, too.