Collectors should employ an organisational asset management system for the audio data that is suited to the type and volume of data and to the objective of the collection effort. The organisational system should meet three needs: the data must be 1) searchable and findable; 2) unique and verifiable; and, 3) reviewable.
First, to ensure data is searchable and findable, data should be effectively labelled. For example, data could be labelled with the time, date, and location of collection or with a unique alphanumeric identification number. Analysts can also assign one or more labels to an audio recording to signal the relevant aspects of the audio’s content. For example, a recording that appears to discuss troop movements could be labelled with ‘possible troop movements’. This allows for audio data to be filtered and thus findable based on its content or by the values of its metadata.
The labelling scheme used by the Collector should be appropriate for the type and volume of data collected. The scheme should be clearly documented, communicated to all personnel, and applied consistently, with training provided if necessary. The Collector should avoid using a labelling scheme where the nomenclature used reaches a maximum value and prohibits further labelling within the same framework. The Collector should also consider whether the labelling scheme is comprehensible to, and interoperable with, foreseeable third-party recipients of the data. The labelling should be included as part of the audio data file’s associated metadata.
The searchability and findability of audio data is further ensured by linking data together. The organisational system should provide for the creation of audio data files, which group together the original copy of the audio data, duplicates (if made), and either include or link to the metadata (see BP 17), such as the records documenting the collection effort (see BP 3). If different audio data files contain mutually corroborating information—for example, if two collection devices recorded the audio at the same time, date, and place—these should be linked within the organisational system. Creating clear linkages between audio data and associated or corroborative data is important for assessing the reliability and authenticity of the data, and thus potentially maximising the data’s probative value in the event that it is presented as evidence in court.
Second, a well-designed organisational system will help ensure audio data is unique and verifiable. Uniqueness is key to preventing accidental overwrites or duplication; verifiability permits demonstration of the data’s chain of custody from the point of collection onward. To achieve this, preservation information generated in accordance with BP 14, such as cryptographic hash values and cryptographic signatures, should be clearly linked with the audio data.
Third, the organisational system must facilitate regular reviews of the data held by the Collector in order to determine whether it is still necessary to retain the data (see BP 2), and whether its retention is in line with privacy protections (see BP 5).1 While it may not be feasible to know the content of every audio recording collected, Collectors should at a minimum have an overview of the data they have collected and stored.
Tech Specs & Resources
For further information on establishing a system that will allow for the verification of the audio data’s integrity and provenance, see e.g., the Coalition for Content Provenance and Authenticity’s C2PA Explainer.
For guidance on the organisation of metadata, drawing from digital audio archiving, see e.g., IASA Technical Committee, Guidelines on the Production and Preservation of Digital Audio Objects (2009), Section 3: ‘Metadata’, as well as e.g., ISO/TC 46/SC11N800R1, Building a metadata schema – where to start.
For guidance on standards-based hashing algorithms and a description of which hashing algorithms are secure, see e.g., Federal Information Processing Standards Publication ‘FIPS PUB 180-4’, as well as e.g., SWGDE Position on the Use of MD5 and SHA1 Hash Algorithms in Digital and Multimedia Forensics (2019).
For guidance on hashing generally, such as what steps a Collector should take when a hash comparison fails, see e.g., NIST IR 8387, Digital Evidence Preservation Considerations for Evidence Handlers (2022), page 7.
The choice of hashing algorithm should be periodically reassessed and revised as needed to take advantage of applicable advances in cryptography.
A number of tools and approaches exist for the production of digital signatures, including the maintenance of a Public Key Infrastructure (PKI), which uses trusted Certificate Authorities (CAs) to issue certificates authenticating identities and securing communications. For more information on PKIs and CAs see e.g., the United States government IDManagement.gov Explainer on the topic.
An alternative to the centralised authorities found in PKI structures are Decentralised Identifiers (DIDs), which are self-sovereign, independent, cryptographically verifiable identities. For more about this novel technology, see e.g., C. Mazzocca et al, A Survey on Decentralized Identifiers and Verifiable Credentials, Arxiv (2024); also see e.g., the World Wide Web Consortium (W3C) Verifiable Credentials Use Cases and their Specification.
Legal Framework
See section 4.3. on the presumption of innocence as an element of the right to a fair trial.
See section 5.2. and section 5.3. on the importance of corroborative information when establishing the relevance and probative value of potential evidence.
See section 5.3. on the role played by chain of custody in the probative value of potential evidence.
Applicable Ethical Principles Accountability; Accuracy, Impartiality, and Objectivity.
Footnotes
-
ECtHR, Guide to the Case-Law of the European Court of Human Rights: Data Protection (2022), para. 205, identifies cases that support the need for regular review of data retention periods and practices, including Gardel v France (ECtHR), Judgment, para. 69 and Peruzzo and Martens v Germany (ECtHR), Decision, paras. 44-49. ↩