Forensic video technicians and analysts utilize various hardware and software tools to acquire, interrogate, interpret, and present image-based evidence in the courtroom. When presenting their evidence in court, technicians and analysts must be able to justify their reliance upon their chosen tools and must be able to articulate why the court should have confidence in their choice. They must also be able to establish that the data discovered or generated by the tool are accurate and reliable. This article discusses the steps that are required to develop this level of confidence.

Tool Validation

Tools are used to perform tasks, and in the image-based evidence context, they are used to discover and generate images and related data that may prove to be useful to investigators, counsel, and the court. Most tools are open source tools, which may be defined as computer programs, software, or other electronic or digital methods whereby data can be discovered, accessed, analyzed, or presented. They may be tools that are only available to someone who pays to use them, or they may be available to the public at no cost.

Tools and the work that they do and the results they produce are not admissible without going through established legal and evidential steps. Further, it is important to distinguish between tools that are suitable only for investigative purposes and tools that can be used for substantive purposes at trial. Tools in the former category can point investigators in the right direction and may be used as a precursor to compile valuable potential evidence, but more corroborative steps are required before any of this data can be used in court. For example, AI may be used for facial recognition analysis at the investigative level, but the results generated by AI are not independently admissible absent corroboration.

Technicians and analysts may rely upon a tool that has already generated images (eg. CCTV, a smartphone, or other camera) or may use a tool to generate images for analytical purposes (eg. reverse projection and other on-scene experiments). Such tools may be privately owned, commercial products, freeware, or open source generally. However, these images and related data cannot be admitted into evidence, let alone trusted, without first validating the tool that was used to produce it. Validating a tool requires the following:

Step 1 – Testimony by a competent, properly qualified witness.

Step 2 – The witness must be fully conversant with the tool being used. The witness need not be an expert in the tool and its use, but the witness must have a strong working knowledge of how the tool works, how to use it, and how to interpret the resultant images and data. If there is training and certification available regarding the use of the tool, the witness should ideally have both completed.

Step 3 – The witness must thoroughly explain how the tool works and what it is intended to be used for.

Step 4 – The tool must be validated by establishing that it correctly and accurately does what it is intended to be used for and was used for by the witness. The witness must be able to answer questions regarding the tool that will generate confidence that the tool is reliable. Validation requires more than just showing how the tool works. It must be established that the tool is trustworthy and fit for purpose.

This level of proof could be provided by a user of the tool, whether testifying in a lay or expert capacity, though tool validation is more soundly situated in the expert witness world since it calls for opinion evidence. Experts use tools daily to accomplish their forensic tasks and must be able to meet the four steps noted above. If they cannot, they should not be using the tool. Note that experts are not required to have a code-level knowledge of the tools that are used as they do not claim to be an expert in the tool itself – when the questions reach that depth, the witness should defer to the tool developer. The developer or representative of the developer may need to be called to testify if the issue is substantive but that would be rare.

Tool Output Data Validation

Tools are used to evaluate existing data or to generate new data. In the image-based evidence realm, this means that selected tools are used to evaluate pre-existing images and are also used to generate new media through image clarification processes, editing, timeline creation, experimentation, etc. When images and other digital evidence are presented, such evidence must be proven to be accurate and reliable – in other words, it must be authenticated. In Canada, this is set out in the Canada Evidence Act, as follows:

S. 31.1 CEA – “Any person seeking to admit an electronic document as evidence has the burden of proving its authenticity by evidence capable of supporting a finding that the electronic document is what it is purported to be.”

S. 31.8 CEA – “Electronic document” means data that is recorded or stored on any medium in or by a computer system or other similar device and that can be read or perceived by a person or a computer system or other similar device. It includes a display, printout or other output of that data. 

In the United States, the authentication requirement is set out in Federal Rule of Evidence 901 as follows:

Rule 901. Authenticating or Identifying Evidence
(a) In General. To satisfy the requirement of authenticating or identifying an item of evidence, the proponent must produce evidence sufficient to support a finding that the item is what the proponent claims it is.

Similar provisions are found in each state evidence code. There is also a common law authentication requirement which applies to all evidence.

Pre-existing and generated images and other data are authenticated by meeting the following requirements.

Step 1 – Authenticate and establish the accuracy and reliability of the underlying data that was pre-existent and that was used by the validated tool to generate further data. This will include a metadata analysis (when metadata is present) and a content-based analysis.

Step 2 – Show that the data has been correctly incorporated into and analyzed by the tool. To accomplish this, the witness must be fully conversant with the selected tool.

Step 3 – Establish the accuracy and reliability of the generated data.

This testimonial data authentication evidence can be provided by witnesses who have sufficient qualifications to validate the tool and the pre-existing and output data. It follows that technicians and analysts who are not fully conversant with the attributes and use of the relevant forensic tool should not use it for forensic work as they would not be suitably positioned to present and defend the evidence in court. Counsel and the court must be able to rely upon properly qualified witnesses to state why a selected tool was a suitable tool to be used in the case and why the pre-existing and output images and other data are accurate and reliable.