Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Facebook denies 'listening' to conversations

bbc.com

Post reply on HN
Print view View original

401–410 of 415 posts

Re: Facebook denies 'listening' to conversations

#401

Post by oh-kumudo »

wu-ikkyu wrote: post #193
oh-kumudo wrote: post #115

From a technical perspective, it is not very likely that they will listen to people's voice to push advertisements. They needs to listen to you ALL the time, collecting your voice from all kinds of noisy environment, uploading to their server for further analysis. And in order for this service to be profitable, they need to train model to extract the part of your conversation that hints a shopping intention. This req…

>in order for this service to be profitable, they need to train model to extract the part of your conversation that hints a shopping intention So basically just listen for any time someone says a variant of "I need/want x"?

Well, there are a lot of needs: I need to go. I need some sleep. I need more time. I want to quit...like you name it.

Re: Facebook denies 'listening' to conversations

#402

Post by goialoq »

sp527 wrote: post #15

Once happened while remarking on the Fruit Loops shake at Burger King. Opened Facebook minutes later and got a Fruit Loops ad. I assure you I don’t message, post, or in any other way remotely suggest an affinity for Fruit Loops. Commenting on it at Burger King was the definition of a one-off incident. I guess another possible explanation is some very sophisticated combination of location tracking and Kellogg’s target…

You think they need microphone access for that? Froot Loops is in the middle of a massive co-branding campaign with Burger King, and you are surprised that you'd get an ad for Froot Loops while at a Burger King? Occam's razor says no.

Re: Facebook denies 'listening' to conversations

#403

Post by goialoq »

codingdave wrote: post #99

The thing that frustrated me with Lance Armstrong over so many years was the smirk on his face when asked whether he used drugs, and he would reply, "Hey, I passed the test, isn't that good enough for people to stop asking?" I suspect Facebook is falling in that same boat - technically following rules so they can answer questions the right way. But we're not seeing the smirks on their faces as they give that answer.

Besides Facebook denial and the complete lack of empirical evidence, what are you looking for? Why do people keep talking about "this one time it maybe happened", when it should be happening to everyone all the time, if it's happening at all?

Re: Facebook denies 'listening' to conversations

#404

Post by sharkmerry »

imron wrote: post #268
null0pointer wrote: post #227

Earlier quoted context omitted.

I have a similar anecdote. Girlfriend was having some symptoms, we briefly talked about it then walked to the pharmacy and bought a pregnancy test. Maybe 30 minutes after purchasing the test we watched a youtube video on her phone. The ad on the video was for the exact test we bought. I figured that youtube got the credit card info from the purchase but I was shocked at how quickly that translated into something show…

Advertisers are suckers. Facebook and YouTube are selling ads to people who have already bought their product.

im bombarded with ads for Ecwid. a product i purchased and have used for 5 years. I feel bad, they must have spent more on ads than I actually pay them.

Re: Facebook denies 'listening' to conversations

#405

Post by sharkmerry »

londons_explore wrote: post #356
ravenstine wrote: post #289

Earlier quoted context omitted.

In a sense, they are suckers. Plus, as another story on HN suggested, their abandonment of advertising you things you don't need right now(but you may find you want later in life) proves counterproductive. A heck of a lot of things advertised to me on Facebook are things I already searched for and either purchased or already decided wasn't worth buying. I don't believe that modern internet ads are as valuable as the…

Internet ads are far more measurable than most other forms of advertising. The advertiser certainly has all the data necessary to see if it's profitable or not to do those seemingly counterintuitive ad campaigns.

except the data is controlled by the entity selling you the ad, so they have reasons to make the profitablity difficult to discern

Re: Facebook denies 'listening' to conversations

#407

Post by crucifiction »

artificial wrote: post #295
crucifiction wrote: post #128

Earlier quoted context omitted.

Its only useful to train if they also have the transcript by a human. Eavesdropping conversations and having armies of people transcribing them seems like a very expensive and illegal way to get that data when there are probably millions of available samples, TV shows, etc that have both voice and transcription available already off the shelf.

The 2017 F8 developer conference featured a lot on Machine Vision. Processing images and video for objects, such as for automatic close captioning, is where vast resources are focused. I highly recommend watching a few of the videos, they’re specially aware as well and can infer orientation of obscured things like limbs. Microsoft has real-time audio translation, doing machine transcription at scale is totally feasib…

The OP that I was replying to was insinuating that FB is collecting audio as training data to create AI models like the ones you are talking about. I was pointing out that raw audio is useless to train an AI model for recognizing words, the whole point of training data for AI is that you have an input and a known output (transcription) that you can use to train and test the model with, having just input is useless for training.

Re: Facebook denies 'listening' to conversations

#408

Post by bearbearbear »

dnate wrote: post #355
bearbearbear wrote: post #274

Earlier quoted context omitted.

Hi, first comment so please be gentle. I'd like to point out that this video doesn't actually provide any evidence that Facebook was listening. They said they talked about cat food all day, then they showed an ad on Facebook for cat food. There's no way to verify that they didn't make the video of the cat food ad on Facebook and then talk about cat food all day. Nor is there any evidence that they talked about cat fo…

That's been done, and no one has found any evidence that this is happening. Hence it's status as a conspiracy theory. As soon as you start assuming that the OS is giving the facebook app access to the microphone regardless of whether you allowed it in the settings, things start to get absurd

Link please?

Re: Facebook denies 'listening' to conversations

#409

Post by Buge »

kagamine wrote: post #359
Buge wrote: post #308

Earlier quoted context omitted.

That could easily have been a random ad, combined with something like the Baader-Meinhof phenomenon. To be a scientific test it needs a list of products, and then randomly assign half into a study group and half in to a control group. Then talk all the time about stuff in the study group but not the control group. And see if there is a statistically significant difference between the ads shown for products in the stu…

IN the video he claims not to have owned a cat or talked about cats for 20 years. they chose a subject that they are certain they have, as a couple, never exchanged words on. Explaining away doesn't explain. For an average, normal FB user, this was as scientific as they could get. It would be nice to see someone repeat the experiment with a pro-hacker type on hand and a packet sniffer ++. There are too many people gi…

Except he probably sees hundreds of ads for things he's never talked about for 20 years. It's not scientific at all, because he has no way of distinguishing a listening ad from luck. My statistical method would distinguish that.

But I guarantee you there are hundreds of pro-hacker types not giving facebook the benefit of the doubt. They're reverse engineering the apps, monitoring API calls from rooted phones, monitoring network traffic, etc. There are entire forums of people dedicated to hacking on android and finding rooting methods, etc.

There are tons of security researchers in academia desperate for a nice paper, eagerly looking for something juicy like facebook app listening to users. In fact these researchers have found that hundreds of android apps are listening to you[1]. But facebook isn't on that list.

The EFF has tons of smart people eager to dig into any little privacy mistake that a major company makes. They for sure would launch a huge lawsuit against facebook if this came out, and they very likely have some researchers looking into whether there's any listening.

Why haven't you heard about all the pro-hacker research into facebook's spying? Because they have all found no spying, because there isn't any. They don't want to publish that they found nothing, because then it would look like they're defending facebook, and these pro-hacker types don't like facebook and don't want to defend them.

[1] https://arstechnica.com/information-technology/2017/05/there...

Re: Facebook denies 'listening' to conversations

#410

Post by sobani »

sharkmerry wrote: post #405
londons_explore wrote: post #356

Earlier quoted context omitted.

Internet ads are far more measurable than most other forms of advertising. The advertiser certainly has all the data necessary to see if it's profitable or not to do those seemingly counterintuitive ad campaigns.

except the data is controlled by the entity selling you the ad, so they have reasons to make the profitablity difficult to discern

But still, when you advertise X to people who searched for X' (or even X itself), you can see the conversion rates of those ads, right? And you should be able to measure if that setup is profitable or not.
Post reply on HN