Mapillary joins Facebook

I compress all my files to 1.5 Mb, since its an archive from the past with limited emotional value with as main target to facilitate mapping and watch my bicycle trips on a map. (Maybe I will never find an alternative to upload these files). They are not family pictures…

Yes, if I were to keep them I’d compress them too I think. However my current plan is to just reupload everything on OSC so I’ll just have them on my pc temporarily.

I am thinking more of an institute as the National Archive or Library or Geographical Institute to donate the pictures to.

2 Likes

That is a good idea actually!

I’ve thought about uploading to archive.org too.
But I assume you’d have to ask them before; currently there are 240k community images, adding hundreds of thousands (or in your case, 12 millions!) of pictures would just flood the image archive.

That gives a “Raw image dump” feeling… there is much added value with the whole thing projected (correctly) on a map…

3 Likes

Ok, my heart’s rate has gone down now, breath in, breath out… I found the original post announcing this, it’s a good read:

What I am missing is a (good & trustworthy) story concerning privacy… I trusted Mapillary with this… I don’t trust Facebook with (my) privacy. In this deal… is there anything there that could change my trust issues in this matter?

Mapillary? @katrin, @eneerhut, @Anders, @jesolem?

When Facebook will be bought by the Chinese, then I will worry.

Right, try to explain this to a city council, or even worse, after you have convinced the city council or a highway authority to contribute to Mapillary…now Facebook. :sweat::fearful:

1 Like

Why? Because American spies are better than Chinese spies? :rofl:

3 Likes

Facebook’s interest in the imagery and data is specifically for mapping purposes and especially for contributing to OpenStreetMap. The main metadata of interest is the location of that image, the time it was taken, and the make and model of the camera. This is all information that Mapillary has collected and utilised to build better tools.

The only difference now is that Facebook is actively reliant upon and improving OpenStreetMap. This means more emphasis on building tools and utilising imagery and data that support this effort.

2 Likes

I see so many problems with OSM that I would not mind if FB started a fresh map.

At least you get to fix things quickly in OSM. Took me years to get blatant errors fixed in Google maps!

2 Likes

What are the problems you see? I think it’s a marvelous project.

3 Likes

Excuse me? Maps should be public domain and not the “property” of one company anywhere. Example, Google shows different maps in different country’s to not upset their (potential) customers/advertisers/users…:
https://qz.com/224821/see-how-borders-change-on-google-maps-depending-on-where-you-view-them/

Things like that are simply horrid, seated one the simple principle: don’t offend, make money. The truth isn’t always easy, don’t hide from it, face it.

Thank you @eneerhut for finally someone from Mapillary joining the conversation on this important topic. :+1: Because many things are still unclear about what the Mapillary-Facebook merger means for the data and users.

Facebook’s interest in the imagery and data is specifically for mapping purposes and especially for contributing to OpenStreetMap. The main metadata of interest is the location of that image, the time it was taken, and the make and model of the camera. This is all information that Mapillary has collected and utilised to build better tools.

So you say. However, we do not know nor can be sure at this point in time that Facebook is not going to use the raw data for any other purposes than “specifically for mapping” some time in the future.

The only difference now is that Facebook is actively reliant upon and improving OpenStreetMap. This means more emphasis on building tools and utilising imagery and data that support this effort.

In fact, this is not true. Facebook could have used the data on Mapillary to contribute to OpenStreetMap without the acquisition of Mapillary, just like any other OpenStreetMap contributor and regardless of the fact what stake OpenStreetMap might have for Facebook. So, your statement is inconsistent. Facebook could have just licensed the data if they felt or knew that there was something they could not do with the data for mapping purposes. In that case, a major licensee like Facebook would have definitely helped with Mapillary’s business model and would have also kept the balance of stake holders, especially for those with no voting power on any board of directors, namely community members and contributors.

Facebook could have also licensed the machine learning technology from Mapillary and developed it further in a primer partnership. Instead, what we got is the “empire firing the Death Star on a pacifist peaceful planet”. Or, since Mapillary wasn’t passive either in this deal, we maybe got more something like “joining Sauron out of madness”.

Imho the community’s major concern with Facebook is that Facebook is simply put unfit to be a good custodian of the data on Mapillary, or rather any data for that matter. Mapillary could/should have been that good custodian and a reliable gate keeper as to for what purpose the crowed sourced data should have been licensed. Now, none of this can happen anymore. Because with Mapillary’s unconsidered and single handed action to merge with Facebook, all the trust which may have existed in the community for Mapillary being a good custodian of the data, has been vaporized. In other words, Mapillary has disproved itself to be a good custodian. All it took for Mapillary was some number X of :heavy_dollar_sign: to throw all of this trust overboard. So, the situation is even worse than most people perhaps realize; personally, I see little to no way for Mapillary/Facebook to restore this trust now.

3 Likes

I see so many problems with OSM that I would not mind if FB started a fresh map.

There is no issue with OpenStreetMap and Facebook. They are a valid contributor, just like any other. And, if you look closely at the ODbL it is extremely permissive. You are allowed to use OpenStreetMap data to build even weapons, i.e. automated drones, cruise missiles, targeting systems, layout mine fields, and the like!

1 Like

@eneerhut, thank you for responding (quickly)!

Normally I would send personal messages with “sensitive subjects”, but I think it is important others can read along. I hope It helps to get (more) people “on the same page”.

Reading your response and the press release of @jesolem again I gather that FB sees value in maps. In trying to understand why, I read that good maps have value for their marketplace platform and I found this video:

That I did not know, that is a way where FB can actually really help. Up until now I have had the distinct impression that FB emphasizes extremes and therefore tends to devide more then it unites. I’m actually impressed that they found a way where FB can genuinely “do good”.

OK, thus, FB sees the benefits of improving maps and see the (greater) value in using & improving OSM, good.

In that goal FB sees the benefits in partnering up with Mapillary… ok, I tend to understand that. But FB “bought” Mapillary, did they not? I have a bunch of questions, but I think I can sum up a load of detail questions by just one word: privacy

My main concern is that I do not want, in any case, ever, what so ever, that imagery I entrust to Mapillary gets abused for face recognition or car/licence plate detection & recognition etc… Combine this with the beyond horrible idea of the combination with time & location…

I understand that Mapillary needs to keep the original image on file in the case of erroneous blurs, I think that I need to hear/read that Mapillary wil not, in no case, ever, with the only exception of a court order share an original image.

The imagery I uploaded contains a face or a license plate here and there. This can not always be avoided. I trusted Mapillary blurring them out correctly and trust Mapillary not (ab)using this data. Now, with this development with FB I read your terms again:

2.2: You hereby waive any moral […] rights you may have (including […] integrity, privacy, […]) in your Content that would otherwise preclude us from […] using, copying or distributing, in accordance with these Terms, your Content.
2.4: You agree not to upload […] infringement of the intellectual property […] or violation of a right of privacy […] or divulges other people’s private or personally identifiable information without their express authorization and permission

When I read the text snippets above, FB (and Mapillary) could do exactly what I don’t want and when someone takes legal action I, the uploader of the image, am the one that’s screwed?

I searched your terms for a part where it states that any personally identifiable sections of uploaded images are obscured/blurred out to safeguard privacy to the best of the ability of Mapillary, but that Mapillary can not be held accountable… or something to that scope… but I could not find it.

As I wrote earlier I trusted Mapillary not to misuse my additions to the Mapillary project. The reputation of FB is not as goed as that of Mapillary… afaik… Is there anything you can devulge cencerning this?

1 Like

Thanks for the detailed reaction, I agree a lot with your reactions and feelings.

Here are a few of my thoughts after reading your reaction:

To be honest, I intensely dislike that “Facebook Data for Good” video. What Facebook is showing there is their ambition to drag both victims and helpers away from already available, often free and open, resources and instead lock them into their services. Every time they mention “Facebook maps” in that video, they are actually talking about OpenStreetMap with their (thin) Facebook skin applied over it. I don’t know if you are familiar with the Humanitarian OpenStreetMap Team (https://www.hotosm.org/), they have long been providing free and open mapping data “for good” in partnership with organisations like the Red Cross and Doctors Without Borders.

As far as privacy is concerned, Facebook and Mapillary are both bound by the same laws (and have the same freedom) in this respect. Yes, you as an uploader are (and have always been) responsible for following the local laws in what you upload. However, with regard to masking license plates and faces, the general laws there are concerned with hosting/publishing of information (that is, Mapillary is responsible), not if you took a photos of it and/or stored that photo somewhere. However, if for example you took a photo of a state secret and uploaded that to Mapillary, then you could well be held responsible.

Reputation and trust are of course very important emotional factors, but I don’t think they make much difference in actuality. Mapillary could well be doing much worse things than Facebook has done, but gotten away with it unnoticed because they are a much smaller company without all the public scrutiny.

2 Likes

I feel it is important to clarify that HOT is great at attracting OSM contributors and motivating them. We are still talking about OSM data here.
Similar as it wouldn’t be good for Facebook to grab all credit for OSM data, it wouldn’t be good for HOT to do so.

1 Like

@eesger Thank you for your hint at Facebook’s involvement in OpenStreetMap.

Facebook Data for Good

This title itself is very telling of Facebook’s mindset and the culture prevalent in this corporation. Because if you take a step back and look at the whole picture, this means that usually Facebook data is used for bad. So, they are well aware that they are doing bad stuff with the data they collect on their users (and others) in everyday business. And apparently, what they are trying to do with that “Facebook Data for Good” initiative, is to free their conscience and tell the public: Hey look, with the data we have collected, we do not only do bad stuff but also little bit of good stuff too! This is like if the mafia was supporting charities with blood money.

The problem they have is that their business model is twisted and malicious in the first place. So, they try hard to grasp any pseudo plausible explanation why they should carry on with it. Businesses, data, and basically anything should always be used for good only, and not just occasionally. Doing bad should be, if ever, a misstep only, unintentional. You should also try to fix your mistakes or missteps. However, they are not inclined nor willing to fix their business model. Instead, they are covering the existing bad business model with a few good examples, which in the end can still serve their initial bad business model. This is indeed twisted.

1 Like