From Digital Evidence to Defensible Conclusions: Rethinking Digital Forensic Investigations with AUTID
Every digital investigation begins with traces. A log entry recording a user login, a suspicious executable found on a compromised machine, an unusual network connection or a series of modified files. Each may provide valuable information about an incident. Yet none of these observations, considered in isolation, necessarily explains what happened. For digital forensic investigators, the challenge extends beyond collecting and analysing digital evidence. They must establish how individual observations relate to one another, determine which conclusions can reasonably be drawn from them and reconstruct events in a way that can be independently verified. This distinction lies at the heart of a fundamental challenge in digital forensics: identifying digital traces is not the same as establishing what those traces mean.
The growing complexity of digital investigations
Modern digital investigations involve an increasingly diverse range of information sources. Investigators may examine disk images, memory captures, operating-system logs, network traffic, cloud environments, mobile devices and threat-intelligence reports, often within the same case. These sources provide information at different levels of detail. A memory capture contains raw digital data. A system log records observable activity. A security alert may already contain an interpretation of suspicious behaviour. A threat-intelligence report may associate that behaviour with a known attack technique. Consider a ransomware investigation. Investigators may identify evidence of a suspicious programme executing, files being modified, unusual authentication activity and connections to external infrastructure. Individually, these observations provide fragments of the incident. Understanding whether they are connected, what activities they represent and how they contributed to the attack requires additional analytical reasoning. The difficulty is not simply the volume of information. It is the need to transform heterogeneous observations into a coherent, evidence-supported understanding of events.
When forensic tools provide answers, but not explanations
Digital forensic investigations rely on specialised technologies to extract, process and analyse information. These tools can recover deleted files, identify suspicious activity, correlate timestamps and classify attacker behaviour. However, the results produced by forensic tools are not automatically equivalent to established investigative conclusions. For example, a security monitoring system might associate suspicious activity with a technique from the MITRE ATT&CK knowledge base. Although this classification provides useful investigative context, it does not, by itself, demonstrate that the technique occurred. The interpretation must be supported by the underlying observations.
A similar challenge arises when different forensic tools analyse the same incident. Their outputs may use different formats, terminology and levels of detail, leaving investigators to establish how individual findings contribute to a broader explanation. When analytical decisions are not explicitly documented, it becomes difficult for another investigator to reconstruct how a conclusion was reached or assess whether an alternative explanation is possible. This challenge becomes particularly significant in criminal investigations, where findings may be examined by other experts, law enforcement authorities, prosecutors and courts.
A forensic conclusion must be more than technically plausible. The reasoning supporting it must also be transparent, traceable and open to independent scrutiny.
The need for a structured approach to forensic reasoning
Established digital forensic methodologies provide important guidance on evidence acquisition, examination, analysis and reporting. However, organising an investigation into procedural steps does not necessarily explain how individual observations should be transformed into higher-level conclusions. How does an investigator move from identifying a registry modification to establishing that a programme was configured to execute automatically? How can several observed actions support the identification of a broader attack technique? And how can those techniques contribute to reconstructing the sequence of events underlying a cybercrime? Answering these questions requires an explicit connection between evidence and interpretation. Investigators need to progress from observations towards conclusions, while retaining the ability to return to the original evidence to verify, challenge or reconsider their findings. Equally important is the ability to accommodate different investigative objectives, ranging from confirming an individual action to reconstructing an entire incident, without depending on a specific forensic tool or analytical workflow.
Introducing AUTID: A framework for reasoning about digital evidence
Within the ENSEMBLE project, AUTID (a Digital Evidence Reasoning Framework) has been developed to address these challenges by providing a structured approach to interpreting digital evidence and reconstructing events. Rather than replacing existing forensic technologies or prescribing a fixed sequence of investigative steps, AUTID focuses on establishing transparent and verifiable connections between digital observations and investigative conclusions. The framework organises forensic reasoning across different levels of abstraction, ranging from raw digital data and identifiable artefacts to observable actions, attack techniques, broader objectives and event reconstruction.
By establishing relationships between these levels, AUTID provides a common structure for combining information from heterogeneous sources and explaining how individual findings contribute to broader interpretations. Its approach supports both the construction of conclusions from available evidence and the examination of those conclusions against their underlying observations. Furthermore, AUTID accommodates different investigative objectives and incorporates considerations relating to evidence relevance, integrity and legal constraints. Through its tool-independent and technology-neutral design, AUTID aims to support consistent, reproducible and traceable reasoning across different investigative environments.
Towards more trustworthy digital investigations
As digital investigations increasingly combine evidence from multiple devices, systems, organisations and jurisdictions, the ability to explain how conclusions are reached becomes essential. Within ENSEMBLE, AUTID contributes to the project’s broader objective of supporting forensically sound cybercrime investigations, particularly in complex cases involving ransomware, unauthorised access, data theft and other cybercriminal activities.
AUTID does not seek to replace the expertise of digital forensic practitioners. Instead, it provides a common reasoning structure that supports the systematic interpretation of digital evidence. Its contribution lies in addressing a fundamental question: not only what can be discovered from digital information, but how investigators can demonstrate that their conclusions are supported by the available evidence. In digital forensics, the strength of an investigation depends not only on the evidence collected, but also on the quality and transparency of the reasoning that connects that evidence to what is claimed to have occurred.
Written by Athanasios Dimitriadis, Efthymios Lyssaris, Dimitris Kavallieros & Ioannis Chalkias from CERTH.
