5 ms·
> (What it actually shows is the results it thinks the person asking the question wants to see, which is wildly different person-to-person, but presented as obj
by creato 8y ago
> (What it actually shows is the results it thinks the person asking the question wants to see, which is wildly different person-to-person, but presented as objective truth to the user).
I don't think this is anywhere near as significant a problem as you suggest. I just tried a few incognito vs. not incognito searches of various conspiracy theories that have become much more visibly popular in the last few years. I saw little significant difference between incognito mode and not, and most of the results are about why the particular conspiracy theory is bullshit.
In contrast, it's a small sample size, but I've seen first hand a small number of friends (and one family member) become much more extremist anti-vaccine crusaders, and the primary vehicle for me seeing this is facebook.
I think youtube is much closer to having these kinds of problems than google search, but youtube is also a lot closer to being "social media", which is the general problem, not facebook in particular. Facebook is just far and away the biggest and most influential.
- cheschire 8y agoI'm interested in hearing if you get different results by using TOR browser instead.
- headmelted 8y agoCouple of things. You'd need to check with Tor for a fair comparison to rule out geographic/workplace affinity. also.. Your starting perspective governs the search term. I'm confident vaccines are safe overall - so I'll search for "are vaccines safe?" (positive connotation). Someone who believes that vaccines are causing brain damage has already formed a contrary opinion, so in all likelihood would be expected to search for "vaccines cause autism" (negative connotation). The question is the same, but it's been presented differently, and Google will respond to the question asked. Believe it or not, I'm not actually blaming Google for this. In essence, it's still trying to answer the question asked, but I think we need to seriously reassess what the heck we're doing and maybe launch public awareness campaigns about how this all works at a fundamental level.
- creato 8y ago> Someone who believes that vaccines are causing brain damage has already formed a contrary opinion, so in all likelihood would be expected to search for "vaccines cause autism" (negative connotation). I understand this, in fact, "vaccines cause autism" was exactly one of my test searches. As for Tor: I don't have it and I'm not going to bother. And if nothing else, my region is a well known hotspot for anti-vaccine idiocy, so I would expect that to bias the test in favor of the filter bubble hypothesis, not against it.
- headmelted 8y ago^^ This makes me feel better, believe it or not. I still think an Internet search / social media PSA campaign is urgently needed.
- creato 8y ago> ^^ This makes me feel better, believe it or not. Frankly, why trust me? Why not try this yourself? I wish no one would trust random commenters on any social media, whether it be facebook or hackernews. I'd rather you not trust me, I only spent 5 minutes on it. For someone such as yourself with such a strong opinion on the topic, I think it is worth spending a bit more time on it. This is one of those rare issues where if you think there's a problem, it's not hard to check for yourself.
- headmelted 8y agoI have, insofar as searching from my own connection in multiple incognito windows - but as we've said that tells me nothing. I'm not going to download Tor because honestly I don't want my ISP thinking I'm up to no good (paranoid and ridiculous as I realize that is). Totally agree otherwise on trust not being a substitute for critical thinking and applied reason, and yes, it's the biggest problem of all of this.