Digital evidence and AI: the new battleground in civil litigation
The question is no longer simply whether digital evidence can be formally admitted in the proceedings. The real debate is whether it withstands rigorous cross-examination. A screenshot may serve as evidence, but it is unlikely to carry the same weight as a documented technical extraction, with preserved metadata, demonstrated traceability, and validation by an expert. Similarly, a transcribed conversation, an electronic file, or an image presented without context may prove insufficient if there is no evidence of how they were obtained, preserved, and introduced into the proceedings.
Spanish procedural law provides a flexible framework for the admission of digital media, recordings, and electronic files. However, formal admission does not equate to evidential value. Evidence may be valid from a procedural perspective and yet have limited value if there are doubts about its authenticity, preservation, or integrity. In an increasingly digitized environment, evidence must not only be presented: it must also be explainable, reconstructable, and defensible.
Artificial intelligence raises the bar. In addition to traditional questions of authorship and integrity, there are now more complex issues: whether the content has been generated or altered by AI, whether an automated system has selected or interpreted relevant data, what sources have been used, what human intervention has occurred, and whether the opposing party can present an effective challenge. Where there is a lack of transparency, bias, or risk of manipulation, technical sophistication does not necessarily bolster evidence; sometimes, it can weaken it.
Therefore, anyone relying on automated analyses, AI tools, or algorithmic reports must bear a greater burden of technical justification. It is not enough to simply provide the result. The system's logic, its sources, limitations, the degree of human oversight, and the safeguards applied must be explained. The authority of the tool cannot replace the proof of its reliability.
In this context, traceability becomes a central element. It must be possible to reconstruct the lifecycle of the evidence: the origin, collection, access, transfers, preservation, metadata, and controls applied. This requirement gives new importance to the chain of custody also in civil and commercial proceedings. The more digital the evidence, the more important it is to establish not only what it demonstrates, but also how it was generated, preserved, and presented to the court.
However, digital evidence is not evaluated in isolation or as a purely technical matter. Its effectiveness also depends on its coherence with the rest of the evidence, the quality of its acquisition, and its ability to withstand attacks from the opposing party. Technology does not replace the court's discretionary judgment, but introduces additional elements that must be weighed in legal terms.
This precaution is especially important when it comes to artificial intelligence systems. Their results cannot be presented as objective truths solely based on the tool's reputation. They may contain hard-to-detect errors, amplify biases, or generate conclusions that seem coherent but are, in fact, invalid. In this realm, reliability cannot be taken for granted: it must be demonstrated.
To this, an inevitable legal dimension is added. Many digital pieces of evidence contain personal data, confidential information, trade secrets, or elements protected by fundamental rights. Their use in a court, especially when AI tools are involved, requires justification of the legal basis, proportionality, and confidentiality safeguards. Otherwise, the evidence can be challenged or even excluded.
For companies, the consequence is direct. Document retention policies, data maps, access criteria, activity logs, and protocols for the use of artificial intelligence are no longer merely operational or regulatory compliance matters. In a digital litigation environment, they determine the availability, admissibility, and persuasive strength of the evidence. The procedural strategy begins long before the trial: it starts with information management.
For litigation, the challenge is to translate technological complexity into legal categories that the court can understand. It is not enough to invoke concepts like metadata, traceability, algorithms, or artificial intelligence. What is crucial is to explain why these elements allow for establishing—or questioning—the authenticity, integrity, and reliability of the evidence.
Fabio Virzi, partner of litigation and arbitration at Ecija.
Continue reading the article here