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

Facebook was in talks with hospitals about a proposal to share data

cnbc.com

Post reply on HN
Print view View original

21–30 of 120 posts

Re: Facebook was in talks with hospitals about a proposal to share data

#21

Post by vinhboy »

marcell wrote: post #6

The pile on continues... (1) This is probably a project by a small "research" group at Facebook. The goal of that group is probably to publish papers in a Psych journal or something like that about how they were able to correlate anonymized medical data with Facebook feed updates. Tech companies have these research groups for prestige, they are not central to the company mission. (2) According to the article, the pro…

It sounds to me like the researcher wanted to use this data to identify at-risk groups and get them preventive care.

I don't mean "identify" as in, de-anonymize them, but to use data to figure who may need help (or advertised to -- same difference really).

This actually sounds like a good idea if that's the purpose.

It's not any different than retailers identifying cohorts they can sell to.

It also sounds very similar to what a lot of AI research does (in my limited understanding of it). Take known samples and use that to identify/predict other things...

Re: Facebook was in talks with hospitals about a proposal to share data

#22

Post by fapjacks »

lovelearning wrote: post #3

Although the medical data itself may be "anonymized", surely FB is in a position to associate that data with actual people, given that they know so much about a person's schedule, location, searches and private messages.

Yeah, it's the ole "encrypted in transit" (implying unencrypted at rest) corporate spin gimmick. The data is "anonymized"... By itself. But if you just so happen to have some kind of reference dataset or index, well...

Re: Facebook was in talks with hospitals about a proposal to share data

#23

Post by apotheothesomai »

craftyguy wrote: post #8
marcell wrote: post #6

The pile on continues... (1) This is probably a project by a small "research" group at Facebook. The goal of that group is probably to publish papers in a Psych journal or something like that about how they were able to correlate anonymized medical data with Facebook feed updates. Tech companies have these research groups for prestige, they are not central to the company mission. (2) According to the article, the pro…

I dislike/disagree with facebook more than most folks, but even I can smell a dead horse being beaten.

For years privacy-minded people have complained about the lack of recognition of privacy issues among the masses. Suddenly, story after story about aspects of Facebook's anti-privacy practices are being read by the public. This will lead to more stories about other data mining entities as well.

The more this keeps up, the more concrete privacy issues will be in the minds of many.

And yet, HN is full of dismissals. I've decided that to many here, the idea of being in an "elite", informed group is more important than the actual issues.

Given that you are not the target audience of these stories, are you really in a position to judge whether they have reached "dead horse" status?

Re: Facebook was in talks with hospitals about a proposal to share data

#24

Post by downandout »

nealdt wrote: post #20
marcell wrote: post #6

The pile on continues... (1) This is probably a project by a small "research" group at Facebook. The goal of that group is probably to publish papers in a Psych journal or something like that about how they were able to correlate anonymized medical data with Facebook feed updates. Tech companies have these research groups for prestige, they are not central to the company mission. (2) According to the article, the pro…

I think this story could appear harmless - or perhaps not. The potential gold mine of information that insurers and the health industry could get from this kind of data could be staggering, particularly when we think about claims or the price of the insurance policy. This is what many of us have been suspecting for some time so this confirms our suspicions. I don't think it's a non-story. I feel that if such research…

Quite telling that it has been 'put on hold'. That in itself is a story.

It's been put on hold because Facebook PR knows that people don't read the details of stories. By and large, they read headlines - and many of those are misleading at best. The issue with CA, for example, wasn't a data breach - the data they had was collected in compliance with Facebook's rules at the time. Yet many headlines and soundbites have used the term "data breach" throughout this incident.

So, when 90% of the population incorrectly believes, based on some soundbites and a couple of headlines, that Trump hired CA to hack into Facebook and steal their data, read their minds, and steal the election, you don't want to go forward with something else that might sound scary in yet another mischaracterized headline or soundbite. If putting this project on pause is a story, that story is that Facebook has a PR department that understands its audience....I don't think it says anything one way or another about this project.

Re: Facebook was in talks with hospitals about a proposal to share data

#25

Post by sveme »

lovelearning wrote: post #3

Although the medical data itself may be "anonymized", surely FB is in a position to associate that data with actual people, given that they know so much about a person's schedule, location, searches and private messages.

Deanonymization of medical data is actually pretty easy if you know a little about your target (age group, height, a set of pre-existing medical conditions limit the set of potential people considerably).

Re: Facebook was in talks with hospitals about a proposal to share data

#26

Post by codedokode »

> Facebook proposed using a common computer science technique called "hashing" to match individuals who existed in both sets

But Facebook has the source data from which hashes were generated so they can alway reverse them.

Furthermore, if you take a hash from name and surname, it can be easily reversed because the set of names and surnames is relatively small.

Re: Facebook was in talks with hospitals about a proposal to share data

#27

Post by codedokode »

StreamBright wrote: post #10
marcell wrote: post #6

The pile on continues... (1) This is probably a project by a small "research" group at Facebook. The goal of that group is probably to publish papers in a Psych journal or something like that about how they were able to correlate anonymized medical data with Facebook feed updates. Tech companies have these research groups for prestige, they are not central to the company mission. (2) According to the article, the pro…

How long do you think it would take to de-anonymize this data for a somebody who specialized in this?

Both parties have hashes and source data. If they are able to match people in their lists, then they can find out who they are. So it is a matter of seconds I think.

Re: Facebook was in talks with hospitals about a proposal to share data

#28

Post by codedokode »

rmykhajliw wrote: post #15

If I were a FB owner or in this research data I'll put all my skill to deanonmaze those HIIPA medical data because it's clearly a golden mine for insurance companies and pharmacy. This's really easy to sell those users and their data, because of high demand.

It would be super easy. Facebook has both source data for a person and the hash that was generated. So they can always reverse the hash.

Re: Facebook was in talks with hospitals about a proposal to share data

#30

Post by d33 »

codedokode wrote: post #26

> Facebook proposed using a common computer science technique called "hashing" to match individuals who existed in both sets But Facebook has the source data from which hashes were generated so they can alway reverse them. Furthermore, if you take a hash from name and surname, it can be easily reversed because the set of names and surnames is relatively small.

Not if you use salt.
Post reply on HN