234
I still don’t think companies serve you ads based on spying through your microphone
(simonwillison.net)
This is a most excellent place for technology news and articles.
Difficult to judge. Could be confirmation bias, as well. Meaning you got ads for flight befores. But you were not paying attention to them at that point. Which changed after your session and now you think these are connected. (Or you looked something up about that location and that kicked it off.)
These are the usual findings in the rare cases people are able to trace it back and they write some article or podcast about it. Mainly confirmation bias. And once you interact with one ad that got you taken aback, you're trapped.
Doesn't rule out other possibilities, though. I guess what I'm trying to say is, this counts more as anecdotal evidence. And we have plenty stories like this. It's not enough to infer anything. More a reminder to investigate some more.
And yes, it's good practice to keep your phone someplace else when you're having protected/confidential conversations. Smartphones are very complex and they certainly have the potential to spy on you. In fact we know a lot of the apps and computer code is meant to analzye your behaviour and transfer that information to third parties.
How many anecdotal stories before it becomes data? If hundreds of people are saying that this happens and there's no other explanation? Thousands? How many things can be written off as "Oh, something you don't understand is happening, even if we can rule out basically everything." ?
With the scientific method and anecdotal evidence: kind of never. It's illegitimate to draw that conclusion, this way.
You got to dig down to the facts. Or we can just tell the fact that a lot of people feel that way. And I mean "confirmation bias" is a very good explanation. We also have thousands of people believe in esoterics, homeopathy etc. The mechanics of psychology are well-understood. And it's kind of the reason why we invented science in the first place. Because we found things aren't always as they seem. And there are a lot of dynamics to factor in.
If we want to get to the truth, we have to do a proper study. I'm not an expert on this, so I don't know if we got to that, yet. I know people have demonstrated this is technically possible. But as far as I'm aware people have also taken apart a few of the major apps like Facebook etc, logged the traffic and couldn't find anything that uses microphone data to do targeted advertising.
Conclusion: It's either not there, or we missed it.
Now I want to do some kind of experiment where I speak things into my phone and see what happens. It still seems too much to be coincidental.