[HN Gopher] The inability to simultaneously verify sentience, lo...
___________________________________________________________________
 
The inability to simultaneously verify sentience, location, and
identity
 
Author : DemiGuru
Score  : 51 points
Date   : 2023-08-11 18:46 UTC (4 hours ago)
 
web link (arxiv.org)
w3m dump (arxiv.org)
 
| dcow wrote:
| I think a lot of people try to do identity things without
| understanding the fundamental nature of the problem they're
| attacking.
| 
| For instance: I'm really really worried that governments are
| going to default into an understanding of digital identity that
| involves ownership of an email address and mobile phone rather
| than ability to sign a document.
| 
| Or: I'm really annoyed that software services and web apps have
| clauses like "you can't use scripts or automation software to
| access our API" when a browser is just _that_. And they should
| really be enforcing rate limits and punishing abusive behavior
| whether the user clicked a button in a browser or a script did.
| 
| These type of things are not rooted in a fundamental
| understanding of identity, they're sloppy stop-gaps. Despite all
| its faults, this is one of the reasons I'm super excited about
| WebAuthN. At least is make common the idea that an identity is a
| cryptographic secret and not a "possession of an email and
| phone". We really really need to dig out of this "email address
| identifies you" hole.
| 
| Anyway it's exciting to see people discuss the topic more
| formally. It gives me hope that we can ultimately get to a better
| understanding of digital identity and not be trying to solve
| impossible problems by chasing an impossibly perfect solution
| that verifies all 3 tenets that actually doesn't exist and making
| a big mess of things because nobody stopped to ask or understand
| the scope of what we _should_ be trying to do.
| 
| Ultimately identity should be empowering not oppressive. And
| right now it feels more like services oppress people into all
| sorts of weird requirements like having an email, getting a phone
| verification code, running software on a device that has an
| integrity attestation framework, etc. rather than trust them and
| punish bad behavior.
| 
| I want my government (and web services, but especially my
| government) to trust me and punish bad behavior, not treat me
| like an untrusted bot that needs to be managed and continuously
| verified.
 
  | OfSanguineFire wrote:
  | > an understanding of digital identity that involves ownership
  | of an email address and mobile phone rather than ability to
  | sign a document.
  | 
  | Buying a SIM card in a great many countries today already
  | requires showing state-approved ID and then signing a form that
  | the shop clerk prints out. So, ownership of a mobile phone
  | number does mean being able to sign a document. Are you
  | concerned about SIM-jacking? I admit, I find the thrust of your
  | post difficult to follow.
 
    | Brian_K_White wrote:
    | It absolutely does not, since a sim is just an object which
    | can be acquired any number of ways than the intended one.
    | They can also be faked and not even possessed at all, since
    | all the server sees is some data. The server (the phone
    | companies hardware) is not a notary public watching you sign
    | something after verifying that your ID matches your person.
    | 
    | This is exactly the grossly naive assumption they are talking
    | about people treating as though it had any substance at all
    | when it has practically none.
 
  | version_five wrote:
  | >Ultimately identity should be empowering not oppressive.
  | 
  | I agree completely. Are there any examples of that though,
  | where it is empowering?
  | 
  | Talking about government (or any bureaucracy) there is no
  | chance of empowerment, at least from the administrative arm.
  | Software is written by lowest bidders for the convenience of
  | administrators, to help them treat people like cattle. This can
  | only change when we complain to real politicians who could
  | potentially advocate for empowerment. As long as bureaucrats
  | are in charge, it only gets worse.
 
  | dontupvoteme wrote:
  | To these people, anonymity is the problem. being easily doxxed
  | and public is considered a _good_ thing to them.
 
| Morizero wrote:
| Title should read "sentience, location, and uniqueness", which
| the paper states are the three key properties of identity
 
| someguy7250 wrote:
| IMO, the meat of this paper is in section 4.3 and 4.4.
| 
| And I cannot say for sure, but the formal proof of 4.4 basically
| summarizes the same points pointed out in 4.3.
| 
| Most of these are not inherently mathematical problems but a
| social one.
| 
| > Verifying sentience is a fuzzy concept. While they can be bound
| together momentarily as we see in [66 ], the binding is very
| easily decoupled.The verified user might choose to sell off their
| uniqueness identifier at time period t + 1 if the verification
| which binds sentience with uniqueness ends at t.
| 
| Basically, people can sell identities
| 
| ----
| 
| What really concerns me though, is how much and how often this
| paper discusses DRM, or in their own words, a "trust anchor"
| 
| > With the assumed threat model in our case, the lack of inherent
| trust in the user only compounds the unreliability of the model
| without any trust anchor.
| 
| > Assuming a proof of location is for a mobile device, rather
| than a particular human being, then associating the proof of
| uniqueness obtained under such a condition, i.e., without the
| involvement of a trust anchor, is unreliable.
| 
| I know that the authors aren't directly calling for more
| centralized trust. But given recent development at Google, we all
| know how the readers would think
 
  | MichaelZuo wrote:
  | > > Verifying sentience is a fuzzy concept. While they can be
  | bound together momentarily as we see in [66 ], the binding is
  | very easily decoupled.The verified user might choose to sell
  | off their uniqueness identifier at time period t + 1 if the
  | verification which binds sentience with uniqueness ends at t.
  | 
  | > Basically, people can sell identities
  | 
  | I can see why Sam Altman believes iris scans are the future,
  | it's definitely much more cumbersome to 'sell off' your iris.
  | Especially if it needs to be rescanned on a daily basis or
  | sooner.
 
| BSEdlMMldESB wrote:
| we cannot live in a society where I must demonstrate to another
| human that I'm human with a piece of paper just because the other
| human is a bureaucrat with a computer.
| 
| this is an 'online only' problem
 
  | chayesfss wrote:
  | [dead]
 
| RcouF1uZ4gsC wrote:
| Nit Pet Peeve: Confusing sapience and sentience.
 
  | Scaevolus wrote:
  | On the internet, the problem is rarely dogs getting up to
  | mischief.
 
    | rocketbop wrote:
    | That's a nice line but could you tell me what it means?
 
      | willturman wrote:
      | I took it as a tongue in cheek reference to this:
      | 
      | https://en.wikipedia.org/wiki/On_the_Internet,_nobody_knows
      | _...
 
      | ketzo wrote:
      | We consider many animals to be "sentient" -- we recognize
      | that they are to some degree conscious, in this case
      | meaning that they have the capability to sense, perceive
      | their world, and have some kind of emotion about the things
      | they perceive
      | 
      | "Sapience" on the other hand, is essentially "human-level
      | intelligence, consciousness, and self-awareness" (from
      | _homo sapiens_ )
 
  | colechristensen wrote:
  | I'm starting to doubt the Internet nitpickers' distinction and
  | definition of the two.
 
    | DougMerritt wrote:
    | Their distinction follows dictionaries that I've looked at,
    | but common usage is clearly diverging from that.
    | 
    | It's unclear to me at what point we should stop saying that
    | lots of people are using words incorrectly, and start saying
    | that lots of dictionaries are sticking with outmoded
    | definitions, but looking at the past, transitions certainly
    | occur sometimes.
 
      | greiskul wrote:
      | There are lots of people that have this view that there is
      | "objectively correct" language, and that you can find it in
      | grammar books and dictionaries. Any linguist worth their
      | salt knows that is a completely outdated and classist way
      | of studying language. Dictionaries are supposed to be
      | updated acording to how people use language, not people be
      | "corrected" to try to follow the dictionary. While language
      | learners speak a language "wrong", any group of native
      | language speakers has rights of ownership of their language
      | as any other group, and as long as they are communicating
      | in a way that they can understand each other, they are
      | never wrong.
 
| version_five wrote:
| > Trolls, bots, and sybils distort online discourse and
| compromise the security of networked platforms.
| 
| In some sense I think the authors' hypothesis is a good thing, ie
| that you can never fully verify someone online. It prevents
| wholesale algorithmic management of people, which is really what
| governments and companies would like to do, and forced some level
| of human contact or at least intervention. I expect it's
| inevitable that they'll find a way to offload the problem onto
| the citizen, for the most part they already have, but I'm
| personally glad it's impossible to assign me some kind of
| infallible identifier that will let me be _The Castle_ style
| abused remotely and without recourse.
 
| dontupvoteme wrote:
| Half the internet thinks troll means "person who disagrees with
| me or (dis)likes thing i (dis)like" - I'm distrustful of people
| who paint them as a big problem on the internet.
| 
| Is this how the western Social Credit system begins?
 
  | dfhanionio wrote:
  | I think we've lost this battle. "Troll" means roughly "person
  | who behaves badly". The word has become useless.
  | 
  | I know language changes over time. This was clearly a change
  | for the worse.
 
  | pphysch wrote:
  | Trolls, as in troll farms, astroturfing, organized influence
  | campaigns, etc. are absolutely a serious problem for any
  | society that pretends to care about democracy.
  | 
  | Especially in the LLM era, where the marginal cost of adding
  | another artificial "voice" approaches $0.
 
    | dontupvoteme wrote:
    | That's not what troll means. Astroturfing as you said is much
    | better.
    | 
    | They're going to be pushing for WEI/Attestation/Requiring
    | easily doxxable accounts using this a bogeyman, aren't they?
 
___________________________________________________________________
(page generated 2023-08-11 23:00 UTC)