And not just when you think it is. If it's in the room, it hears ya!
It all began with a car crash.
I was doing some ironing when my mum came in to tell me that a family friend had been killed in a road accident in Thailand.
My phone was on the worktop behind me.
But the next time I used the search engine on it, up popped the name of our friend, and the words, "Motorbike accident, Thailand" and the year in the suggested text below the search box.
One friend complained to her boyfriend about a migraine, her first ever, only to find the next day she was being followed on Twitter by a migraine support group.
Another had an in-depth chat with her sister about a tax issue, and the next day was served up a Facebook advert from tax experts offering advice on that exact issue.
One reporter mentioned his male colleague seeing online adverts for sanitary pads after discussing periods with his wife in the car.
But surely if the microphone was activated and the handset was sending data, battery life would be even worse than it is now and individual data usage would be through the roof?Tech challenge
I challenged cybersecurity expert Ken Munro and his colleague David Lodge from Pen Test Partners to see whether it was physically possible for an app to snoop in this way.
Could something "listen in" at will without it being obvious?
"I wasn't convinced at first, it all seemed a bit anecdotal," admitted Mr Munro.
However, to our collective surprise, the answer was a resounding yes.
They created a prototype app, we started chatting in the vicinity of the phone it was on and watched our words appear on a laptop screen nearby.
The whole thing took a couple of days to build.
It wasn't perfect but it was practically in real time and certainly able to identify most keywords. The battery drain during our experiments was minimal and, using wi-fi, there was no data plan spike.
"We re-used a lot of code that's already out there," said David Lodge.
"Certainly the user wouldn't realise what was happening. As for Apple and Google - they could see it, they could find it and they could stop it. But it is pretty easy to create."
"I'm not so cynical now," said Ken Munro.
"We have proved it can be done, it works, we've done it. Does it happen? Probably."
Sure, Apple, Google and Facemask (err, I mean "book") deny doing anything of the sort and the article went on to quote a guy waxing philosophical about "coincidences," but the capabilities of our computers and software have developed to a point where if one can imagine a plausible scenario for a novel form of surveillance then it almost certainly already exists and in active use (and this particular development was easily proven plausible, and relatively simple to accomplish).
And as anyone who compares what's reported by mainstream news sources to the actual timeline of certain developments they report on knows, the information such sources provide will always be behind the curve. Sometimes way behind (e.g. see the documentary Citizen Four). So the story posted above may already be (and probably is) obsolete.
It's probably prudent to take stock of where things stand with surveillance and privacy-eroding technology using the most current information you can find, extrapolate from those findings to project 2 or 3 "developments" ahead to get an idea of where things might stand a couple of years from now, then proceed to conduct yourself as if they are already in active use (because they probably are).