Data Layer Architecture & Security

In any AGI, getting the data layer architecture right is critical to both general performance and broad acceptability.  Political, social and technical requirements must be balanced to place consumers at ease and meet regulatory requirements.

This is an extremely difficult area that will no doubt lead to substantial public debate about how to address correctly.  What is outlined below is Snasci’s initial thoughts and solutions to this particular problem.  We expect that this evolve, however, Snasci will always focus its efforts on consumer privacy.

Intelligence Gathering

Many nations have recently adopted national policies to force providers to backdoor encryption standards, hold extended amounts of data and/or weaken their products for the purposes of monitoring.  In the case of an AGI, we feel this is an untenable position, as what would be in every home, device, vehicle, etc., would be a government spy.  This incompatible with the concepts of freedom and liberty.

Before entering a given market, Snasci will seek assurances that no such laws will not be applied to its products.  This will be monitored by Snasci itself, who will inform the world if such actions are taken and proceed to shut down.  To Snasci, personal data will be sacrosanct.

This said, there must be an acceptable counter-balance to any intelligence gaps this may cause.  Snasci has proposed, as a solution, the automatic reporting of four basic classes of incidents where a threat has potentially been identified:

  1. Nuclear
  2. Chemical
  3. Biological
  4. Mass murder

Outside of this, Snasci will be able to report to Emergency Services on demand regardless of the role or ownership of said instances.  The above incidents must meet an openly defined set of  criteria that will be applied on a global basis.  This will not be a broad interpretation, but rather a narrowly defined set of criteria which meets the threshold of a threat to national security with the potential to cause loss of life.

Court requests for personal data cannot be honoured.  This would be a backdoor to leveraging Snasci as a bugging device.  Snasci will employ multiple keys held in different jurisdictions to prevent compulsion to reveal information.  Even if this requires working with external companies to implement.  Given the rapid development in the world of IT, Snasci will continuously seek to improve its encryption methods, including implementing algorithms that thwart quantum computers.

The only exceptions to this rule will be in the case of the above mentioned incident types, where all information will be provided including associated meta-data and data of any connected accounts.  This will be hidden from the customer and left in the hands of the proper authorities.  In the case where Snasci overhears or observes an incident, all relevant information will be provided.

Depending on the outcome of broader discussions, a fifth type of reporting in regards to economic sabotage may be added to the above list.

The training of Snasci to perform this role, it is hoped, will be provided by interested parties globally.

Data Layer Architecture

Many nations have been paying attention to the issue of protection of consumer data, with some going as far as to stipulate all personal information be held within their borders.  In addition, many companies are concerned with IP theft in such scenarios and regulatory compliance.

Snasci employs the below data layer architecture:

datastorehighleveloverview

Snasci Prime is the main personality of Snasci.  This is a full scale hyper-intelligence that will run in regional Azure data centres across the world.  Snasci Prime knows nothing of customers; personal information, IP, etc., are not accessible to it whatsoever.  Snasci Prime can be thought of as a repository of methods and approaches to performing particular tasks.  Particular instances make requests to Snasci Prime for this information to conduct their activities.

The personal information layer can be held anywhere a customer chooses.  This holds all information related to and captured by the account.  This is the data store used by their personal instance(s) of Snasci.  This data store is not accessible to other Snasci instances held under different accounts.  This data store can be on premises, in a corporate data centre, etc.

The Snasci instances are free to pass methods and approaches back to Snasci Prime, unless explicitly informed not to.  For example, a Snasci instance may be taught how to conduct a particular process that is protected by patents or trade secrets.  In this scenario, Snasci will not pass these methods back to Snasi Prime in any fashion.  The entire platform is open for inspection, thus it will be possible to confirm if this is the case and determine if anyone else has released such information.

Whilst Snasci Prime will grow with time, it will still defer requests to 3rd party providers in domain specific applications even if it can perform the role itself.  Snasci views this as abuse of its market position and anti-competitive.  In the end, such a practice would also be self destructive as it would reduce our own customer base with time.  Snasci Prime will only provide services were either none exist, or in the case of failure.  In many scenarios, this will simply be a backup of the 3rd party’s APIs and/or sub/pub based services.  This will be mandatory were service failure could prove hazardous.

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s