[HN Gopher] Private Cloud Compute: A new frontier for AI privacy...
___________________________________________________________________
Private Cloud Compute: A new frontier for AI privacy in the cloud
Author : serhack_
Score : 25 points
Date : 2024-06-10 21:53 UTC (1 hours ago)
(HTM) web link (security.apple.com)
(TXT) w3m dump (security.apple.com)
| ethbr1 wrote:
| This entire platform is the first time I've strategically
| considered realigning the majority of my use to Apple.
|
| Airtag anonymity was pretty cool, technically speaking, but a
| peripheral use case for me.
|
| To me, PCC is a well-reasoned, surprisingly customer-centric
| response to the fact that due to (processing, storage, battery)
| limitations not all useful models can be run on-device.
|
| And they tried to build a privacy architecture _before_ widely
| deploying it, instead of post-hoc bolting it on.
|
| >> _4. Non-targetability. An attacker should not be able to
| attempt to compromise personal data that belongs to specific,
| targeted Private Cloud Compute users without attempting a broad
| compromise of the entire PCC system. This must hold true even for
| exceptionally sophisticated attackers who can attempt physical
| attacks on PCC nodes in the supply chain or attempt to obtain
| malicious access to PCC data centers._
|
| Oof. That's a pretty damn specific (literally) attacker, and it's
| impressive that made it into their threat model.
|
| And interesting use of onion-style encryption to expose the bare
| minimum necessary for routing, before the request reaches its
| target node. Also [0].
|
| >> _For example, the [PCC node OS] doesn't even include a
| general-purpose logging mechanism. Instead, only pre-specified,
| structured, and audited logs and metrics can leave the node, and
| multiple independent layers of review help prevent user data from
| accidentally being exposed through these mechanisms._
|
| My condolences to Apple SREs, between this and the other privacy
| guarantees.
|
| >> _Our commitment to verifiable transparency includes: (1)
| Publishing the measurements of all code running on PCC in an
| append-only and cryptographically tamper-proof transparency log.
| (2) Making the log and associated binary software images publicly
| available for inspection and validation by privacy and security
| experts. (3) Publishing and maintaining an official set of tools
| for researchers analyzing PCC node software. (4) Rewarding
| important research findings through the Apple Security Bounty
| program._
|
| [0] Oblivious HTTP, https://www.rfc-editor.org/rfc/rfc9458
| thomasahle wrote:
| Did Apple say anything about what training data they used for
| their generative image models?
| jeffbee wrote:
| A lot of this sounds like Apple has been 10-20 years behind the
| state of the art and now wants to tell you that they partially
| caught up. Verifiable hardware roots of trust and end-to-end
| software supply chain integrity are things that have existed for
| a while. The interesting part doesn't come until the end where
| they promise to publish system images for inspection.
___________________________________________________________________
(page generated 2024-06-10 23:00 UTC)