0

I am in the process of building a web-based framework to implement a search engine based on Facial Recognition, it will feature a paid service and a members only area, to protect not only the users, but also the ones who have been searched through the crawler. I have a 2 part question and was hoping it could be answered here.

First question being, the image of the subject in mind will be uploaded to the search engine, and it will compress that image down to the face, and hash that image. Once that image is hashed and discarded, it will use that hash to verify the identity of that person, using websites that are commonly used such as Facebook, Twitter, and Tumblr, once identity is confirmed it will crawl known and common dating websites. My question is, Is this first part of data collection against any type of law, because the crawler will also extract not only the person's identity, but also if they are married, and to whom they are married to.

Which leads me to my second question, in the event that no laws are broken, what should the Terms of Service and legal disclaimer cover, to keep any type of legal recourse from coming back onto me as the owner/webmaster of the site.

2

...using websites that are commonly used such as Facebook, Twitter, and Tumblr, once identity is confirmed it will crawl known and common dating websites.

Read the TOS for each of the sites your crawler or bot will access. I'm sure most of them will state that they do not allow such automated access to harvest their content.

Who owns the content (photos, information, etc.) on each site or can legally license the content - depending on each service - may not be the most relevant point in your case; I'm not clear on if you will download images from each site and compare hashes, or not download images, or what.

But your proposed automated access is a big part of the issue. If the TOS says no automated access is allowed, then automated access by you and your software could get you in legal trouble. The penalties for breaking their TOS would be civil in nature; each service could take you to court individually.

You may be able to enter into a licensing agreement with each service to be able to crawl the content under their TOS; that's up to you and a lawyer.

As to your second question:

...what should the Terms of Service and legal disclaimer cover...

Talk to a lawyer to draw up a TOS for your software and service; Law SE is not for legal advice, as per the Law SE TOS: https://law.stackexchange.com/help/disclaimer

And, in fact, if you are

in the process of building a web-based framework to implement a search engine based on Facial Recognition...

then you should talk to a lawyer before you go to far down the development path with an App that could possibly be a legal liability.

  • I saw the disclaimer after I posted the question. Sorry, but I do thank you for the advice. Now, I will not be storing the images that are scraped. They will be hashed down to a face, and discarded after confirmation/rejection. – user13944 Oct 16 '17 at 4:57
  • By downloading the images to your server for your app to apply a hash algorithm you have already stored them, even if you only store them for a few seconds at a time – Shazamo Morebucks Oct 16 '17 at 10:57
  • @ShazamoMorebucks There are many who believe that the public posting of image or text or other copyrightable works automatically grant the limited right for anyone else to make a copy to view, if not also store, unless otherwise stated (e.g., YouTube is primarily a live streaming website where "downloading" is illegal unless otherwise noted). In other words, it's not an infringement because you have a limited license, if that's what you're talking about. The terms of service may then define the limits. – Upnorth Oct 16 '17 at 23:02
  • I understand. I'll reach out to a lawyer about this issue, and clear it up with the webmasters. Thank you for the advice. And the main public facing image is what I was referring to, when I started this project it seemed like a grey area in my opinion, but you both bring up good points. Thank you. – user13944 Oct 17 '17 at 0:47