AI Reads Privacy Policies So You Don't Have To -- and It's Actually Pretty Good (thenextweb.com) 35
An anonymous reader shares a report: Don't you absolutely hate how dense and confusing privacy policies are? Considering they're full of gotchas and intentionally obscure legalese, it's no surprise that hardly anyone bothers to even read them -- we've simply accepted we're giving up our data, and with it, our sense of privacy. But thanks to this new policy-reading AI, things won't have to be this way for much longer. Guard is a recurrent neural network-based app that reads and analyzes privacy terms, so you don't have to. While it can't yet examine policies on request, the AI has rated the privacy terms of a slew of popular services like Netflix, Instagram, Twitter, and Tinder.
What's particularly nifty about Guard is that it not only gives an overall score, but also breaks down the main threats included in privacy policies. It also lists the total number of threats, and looks at past privacy scandals a service has been involved in. Like I already said, Guard has only analyzed the privacy policies of a small batch of popular services, but you can easily suggest new apps for consideration. The app was built by developer Javi Rameerez, who created the software as part of his thesis on AI and natural language processing.
What's particularly nifty about Guard is that it not only gives an overall score, but also breaks down the main threats included in privacy policies. It also lists the total number of threats, and looks at past privacy scandals a service has been involved in. Like I already said, Guard has only analyzed the privacy policies of a small batch of popular services, but you can easily suggest new apps for consideration. The app was built by developer Javi Rameerez, who created the software as part of his thesis on AI and natural language processing.
So... (Score:3)
Re: (Score:2)
it doesn't seem like it would need any of your information in the first place.
Re: (Score:2)
Re: So... (Score:2)
It can't be good. By definition. (Score:1)
Since it can't know what I want or prefer or find OK.
All it can do, is to Apple it: Pick the dumbest common denominator, and expect the user to be that, until he is.
I don't know why, but that pretty much always turns out to be the most opposite of what I am that is practically possible. And usually far more so, than I could even imagine in my most dystopic predictions.
I suspect though, that that is actually true for most people. As the maximum overlap of a large set of sets, is bound to be very small. Which
Re: (Score:2)
You didn't read the summary. It produces, in addition to a score, a breakdown of threats. You can weight those threats however you like.
Rival (Score:2)
Here's my open source TOS-interpreting AI:
10 PRINT "WE OWN ALL YOU DATAZ LOL"
20 GOTO 10
Re: Rival (Score:2)
I'll make a modest addition:
15 SEND(ALL_YOUR_DATAZ, GESTAPO)
This is how AI extinction starts (Score:2)
AI after reading all these privacy policies will inevitably conclude that humanity is beyond any redemption, and nuking it from the orbit is the only way to be sure.
AI can't be any good. (Score:2)
Unless it tells you it's bullshit. You can't enforce a "privacy policy". First off, the government won't let you. And second, it's impossible to verify they don't harvest data and hide it in the basement.
Re: (Score:2)
Unless it tells you it's bullshit. You can't enforce a "privacy policy". First off, the government won't let you. And second, it's impossible to verify they don't harvest data and hide it in the basement.
It's not just privacy but also copyright, like when they tell you that they own every word and photo that you upload to their site and can use them for commercial purposes in perpetuity without your knowledge or consent, and without payment.
Um... no. (Score:4, Interesting)
Don't you absolutely hate how dense and confusing privacy policies are?
Um... no?
Considering they're full of gotchas and intentionally obscure legalese
95% of which is actually necessary in any contract to make it actually stand up under legal scrutiny.
While it can't yet examine policies on request, the AI has...
...performed no differently than any number of pundits writing articles about the latest horrible changes to any given service's policies.
What's particularly nifty about Guard is that it not only gives an overall score, but also breaks down the main threats included in privacy policies. It also lists the total number of threats, and looks at past privacy scandals a service has been involved in.
Great. Now we've turned legal contracts into a game. It's a score to be manipulated like any other. Perhaps this one company only shares your data with one subsidiary. That looks great, unless you realize that it's that company that sells your data to anybody. Any particular gaming mechanism will work until the developers notice, or until the developers hear about a "scandal". Not a lawsuit or change in legislation, but a public scandal fueled by a media frenzy and a popularity contest.
In short, your legal advice will be driven not by law, but by who can manipulate the media the best.
Re: (Score:3)
Great. Now we've turned legal contracts into a game.
EULA contracts have always been a game. [wikipedia.org]
Re: (Score:3)
Re:Um... no. (Score:5, Insightful)
Considering they're full of gotchas and intentionally obscure legalese
95% of which is actually necessary in any contract to make it actually stand up under legal scrutiny.
It wouldn't be necessary if they weren't trying to use said contract to screw you over.
The MIT license, for example, only has about 15 lines of text.
oblig. (Score:2)
"95% of which is actually necessary in any contract to make it actually stand up under legal scrutiny."
The more corrupt the state, the more numerous the laws.
This is genius! (Score:2)
For years I have wanted to start a non-profit that provides a web site where you can look up EULAs and TOSs to see, in a uniform manner, what they state. I really wanted it to have a "jerky license of the month" that would highlight bad licenses. But the hard part is getting a pro-bono lawyer to review them all. The software license analyzer might make this possible! It looks like Guard does almost exactly what I wanted
Re: (Score:2)
What do you need a lawyer for? Did you not learn to read?
Re: (Score:2)
Are you under the impression that the legal meaning of words is typically the same as the meaning in general English conversation?
Re: (Score:2)
What do you need a lawyer for?
I recommend against a web site giving legal advice with no input from a lawyer.
Did you not learn to read?
I'm blind you insensitive clod! Those screen readers are terrible at reading licenses!
YOU ARE: (Score:2)
Well above average at spotting privacy friendliness
Re: (Score:2)
oh yeah , if you want to help train the machine you can do so here :
https://useguard.com/test [useguard.com]
How do we know it was an AI (Score:1)
Who is to say an individual didn't break it down?
And it is a limited list.
Doesn't Solve the problem !! (Score:2)
Great - It reads Netflix , Apple TV, and Amazon Prime Streaming policies. What if they all say "we totally sell everything and do background credit checks on you monthly" and I don't like it?! No TV for me!?
These EULAs and Privacy policies are written for The Company. Not for the consumer. You accept them or don't use the service. A Very binary choice. The entry onto the internet for everyone is walled by the very companies offering the services, who also wrote the policy.
The consumer is not prot
Re: (Score:2)
Well, actually that is not true at all. You can "strike out" their terms & conditions and replace them with your own, and the same with their "privacy policy". And then you offer them money and make acceptance of the money and provision of the service acceptance of the terms of the modified policies. Then, if they take your money they have agreed to the "shrink wrap" conditions you attached to that act. And you have proof that they accepted your terms -- they took your money.
Is not the bidirectional
Re: (Score:2)
What did you think privacy policies
Analyzing privacy policies (Score:2)
Good news! I've run updated the Privacy Policy through the symbolic analyzer and it condensed from 10 pages to less than half a page.
Expectations of consumer privacy: None!
Data to be sold to third parties: Anything we can get our hands on!
Useless tool (Score:2)
Everybody knows EULAs and "privacy" policies essentially require you to surrender all your rights to everything, and sacrifice your firstborn if you have more than one child, in order to use the product or service.
No need to read them: just assume whichever company is behind the product is hell-bent on shafting you any which way they can and take adequate protections to curb their appetite for data they shouldn't have when using their product/service.
Re: (Score:2)
Sometimes you may even agree to have your mouth sewn to the anus of someone else.
CCPA FTW (Score:2)
Wrong solution (Score:2)
So instead of passing laws to require policies to be written in such a way that they be understood by most people, we're now using A.I. to summarize in plain english what the policies contain.
Re: (Score:2)
So instead of passing laws to require policies to be written in such a way that they be understood by most people, we're now using A.I. to summarize in plain english what the policies contain.
So you are saying the developer had the ability to unilaterally change the laws in every country?
Yes, there may be a better solution, but that does not mean that working towards solutions in your grasp is a bad thing. If you only accept 100% perfect you'll never improve.
Take your username to heart - they did something which is a step in the right direction and within their ability. Celebrate that, don't dismiss their efforts because it's not the sweeping solution you want.
hurrah (Score:2)
I, for one, welcome our new AI overlords (Big Brother) who will read all the privacy (exfiltration) policies and decide how much privacy I will have (none) and how much data the advertising (social credit score) bots will slurp (all of it).
flawed thesis (Score:1)