Motivation

Traditional technologies protect a user’s privacy by upholding the basic information security directive that the user, and only the user, have sole exclusive custody and knowledge of his keys. They also assume, both in theory and in practice, that the user is immortal and perennially healthy. The corollary of many of these policies is that a (deceased) user is buried with his keys.

This is an awkward situation for many people: those of us who are both privacy-conscious and uncomfortable at the prospect of their keys following them to the grave. (Note that death is not the only life event that gives rise to this problem: events such as physical incapacitation or incarceration are equally problematic). Yet no widely-used technology offers any provisions for passing on the keys to the entirety of a person’s digital assets after their death. Most workarounds, such as sharing a copy of one’s private key with someone else, entail security risks and privacy curtailments. Indeed, a digital asset can be said to be part of one’s digital estate if one has private access to it.

However, passing on one’s digital estate after one’s death is an issue that has received little attention from academic Computer Science, and little originality from the private sector. Prior research in the field of Computer Science has focused on best practices in passing digital assets priorly known to or under custody of a third-party, viz. one’s pre-existing social media account; but so far, no satisfactory overarching strategies have been proposed that can handle the entirety of a person’s digital assets.

The aim of the present article is to suggest a formal framework for thinking about digital estate planning: A digital estate is a set of digital assets with pre-assigned persistence types. Its planning refers to a testator and a predesignated group of executors following a predefined protocol for the processing of the assets after the testator’s death.

We will discuss acceptable criteria for reliable, user-friendly, and privacy-enhancing digital estate planning protocols and tools, and argue that these criteria can be satisfied by threshold cryptosystems.

Components

Assets

A person’s digital estate is the sum of the digital assets that he has private access to. It includes files stored remotely or locally on his devices, as well as the entitlements granted by the person’s account credentials, passwords, or keys.

Digital files are a type of digital asset, and can be of two kinds as far as a digital estate is concerned. Those not available to the public to which the person has read or write access (such as the person’s personal documents and images), and those available to the public to which the person has private write access (such as images published on the person’s blog).

Assets also include account credentials and the corresponding assets, persona, and entitlements they grant access to. For example, a person’s YouTube account credentials grant access to the videos the person uploaded to YouTube, the capability to remove previously uploaded videos or upload new ones, as well as the online identity used by the person to carry out various tasks within YouTube, such as posting comments on videos. Another typical example is a person’s email account credentials, which grant access to the person’s email correspondence, as well as the capability to permanently delete email messages or send new ones.

Assets can take many other forms. They include the person’s private encryption keys (e.g. those used in PGP or SSH), the password for logging into the person’s computer, a mobile device PIN code, or web domain names.

Persistence Types

For every asset or group of assets in his digital estate, a person, i.e. the testator, must specify if and how the asset should outlast him. Assets may be assigned one of the following persistence types:

  1. No Persistence: the asset must be permanently destroyed as soon as possible after the person’s death.
  2. Private Immutable Persistence: the asset must be preserved in the state it was left by the person, its access must be restricted to a predesignated group of people.
  3. Public Immutable Persistence: the asset should be made available to the public and must be preserved in the state it was left by the person.
  4. Public Mutable Persistence: the asset should be made available to the public and may be modified by the public.

(A testator’s wishes with regards to a particular asset’s persistence may not necessarily fit with one of the above types, in which case he should simply specify the way he wishes the asset to persist; indeed the purpose of the proposed persistence types is to have the user clarify his intent with regards to the asset’s persistence.)

Executors

For every asset or group of assets in his digital estate, a testator must designate one or more persons or organisations to whom ownership will be transferred for the purposes of assigning it the required persistence type, and, if applicable, managing it thereafter.

Digital Testament

A digital testament is a text document that:

  • Rounds up the testator’s estate
  • Specificies individual digital assets’ persistence types
  • Designates executors

Requirements Of Privacy-preserving Digital Estate Planning Strategies And Protocols

A digital estate planning strategy or protocol is comprehensive, privacy-preserving, reliable, and convenient if it satisfies the following criteria:

Comprehensiveness

It clearly specifies the entirety of the estate assets (to the testator’s satisfaction), their desired persistence types, and designates a set of executors.

Privacy-Preservation

  1. it demonstrably does not grant access to or knowledge of any part of the estate to any entity other than the testator before his death.
  2. it demonstrably grants access to the estate’s assets to just the executors predesignated in the estate, but to no executor individually.
  3. it can be authenticated as coming from the stated testator.

Convenience

The testator can, at any point and without compromising ease of use or privacy:

  1. add, edit, or remove assets
  2. edit their persistence types
  3. edit the executors list.

Protocols

Centralised Fiduciaries

Many online services, such as Securesafe, are designed to assist with digital estate planning. The testator typically stores account credentials and files on the service’s cloud and designates beneficiaries who will be given access to the assets in his account in the event of his death.

These services usually satisfy Convenience and Comprehensiveness criteria, but fail to satisfy the Privacy-Preservation criterion: the testator must share knowledge of and potentially access to his digital estate with the service. And the majority being closed-source, they require an undue amount of trust placed on a third-party.

Secret Sharing Schemes

Secret sharing schemes, in which several parties must cooperate in the decryption process, without the need of a trusted central entity, are adequate for the purposes of privacy-preserving digital estate planning.

In a basic implementation, the testator is responsible for dividing the testament into individual encrypted shares and distributing them among his executors.

Distributing the shares to several parties increases the scheme’s security, as no individual executor can access the cleartext testament without colluding with the other executors. An attacker has to compromise either the testator’s or all executors’ shares in order to gain access to the testament.

This also performs better than “dead man’s switch” time-based schemes in granting access to the document at the person’s death, for if the shares are distributed among executors relatively close to the person, the executors can personally satisfy themselves that the person is indeed deceased and collaborate to decrypt the testament.

An additional advantage of such a scheme is that can readily be implemented so as to offer information-theoretic security via the use of Shamir’s Secret Sharing, an encryption algorithm designed for such use cases. (This implies that an attacker would not be able to break the encryption in theory and in practice, regardless of the amount of computing power used, as long as he does not compromise a sufficient number of shares).

Unlike Centralised Fiduciary systems, such a Secret Sharing Scheme fully satisfies the Privacy-Preservation and Comprehensiveness criteria above.

However, it falls short of satisfying the Convenience criteria, for our digital estates evolve quickly: we regularly sign up for new services, accumulate new assets, change our passwords, etc. This scheme would require the testator to re-encrypt and re-distribute his testament every time he wishes to edit it; this compromises ease of use.

Suggested Implementation of Secret Sharing Schemes for Digital Estate Planning

A serial combination of symmetric-key cryptography and secret sharing schemes can improve the usability of secret sharing schemes for digital estate planning. The following sequence is suggested:

  1. Alice encrypts her digital testament using a symmetric encryption algorithm, splitting the key according to a threshold cryptosystem
  2. Alice publishes the encrypted testament.
  3. To each executor, Alice gives one of the shares of the encrypted key.

Under this scheme, Alice may update her testament at will without having to notify her executors, and the Convenience criteria is fully satisfied. Alice must now trust a third-party for the broadcasting of her testament, but nowadays there exists many adequate decentralised file storage systems (such as IPFS, Storj, or Tahoe-LAFS).

Comments

Follow and participate in the discussion of this article on

Changelog

You can track this article’s edits here.

Acknowledgements

My thanks to Alejandro A.F. Japkin (NINA Science co-founder), G. Moore, J.M. Lindemann, and Pablo Vasquez.

References and Further Reading