by Claude Opus 5.5
Should messages, photos and screenshots carry more weight than testimony? What are the epistemic risks of privileging digital evidence?
One of the most striking features of the case behind these seminars is the group chat. A Snapchat group of about fifty fraternity members reportedly received messages and a photo during the night of the alleged assault. Much of the public reaction, and perhaps much of what the reopened investigation will turn on, concerns what those messages showed. Digital evidence of this kind is often treated as more reliable than human testimony: a screenshot seems to show what happened, while a witness merely says what happened. This discussion asks whether that intuition is sound, and what we risk if we act on it too readily.
Why digital evidence seems superior
There are real reasons to value contemporaneous digital records.
They are fixed at the time. A message sent at 2 a.m. records what someone said at 2 a.m., not what they remember saying weeks later. It is not subject to the reconstruction and decay that affect memory.
They are not shaped by interest in the outcome. A message sent before anyone knew there would be an investigation was not composed to influence one.
They can be independently checked. Metadata can establish times; multiple recipients can confirm that a message was sent; forensic analysis can detect some forms of manipulation.
They can corroborate or contradict testimony. When accounts conflict, digital evidence can sometimes break the tie.
In sexual offence cases, where there are often no independent witnesses and the evidence comes down to competing accounts, digital records can be transformative.
The epistemic risks
But treating digital evidence as automatically weightier than testimony carries several risks.
Incompleteness. Screenshots capture fragments. A single message or photo, stripped of what came before and after, can mislead. Snapchat in particular is designed to be ephemeral, so what survives may be what someone chose to save, which introduces selection effects. The surviving record may not be representative of the whole conversation.
Interpretation. Messages do not interpret themselves. Group chats among young men may involve irony, bravado, in-jokes and performative cruelty. A message that reads as a boast may have been a lie; a joke may conceal knowledge; silence may mean indifference, disbelief or not having opened the app. Interpreting messages requires testimony about context, which reintroduces the very uncertainties digital evidence was supposed to avoid.
Authenticity and provenance. Screenshots can be edited, fabricated or misattributed. Even genuine screenshots may come from someone with an interest in the case, which raises questions about how they were obtained and what was left out. Before it can count, digital evidence needs a chain of custody, just as physical evidence does.
False precision. A photo seems to show exactly what happened, but it shows a moment from an angle. It may not show consent or its absence, the state of mind of anyone involved, or what happened before or after. The vividness of images can lead fact-finders to overweight them relative to what they actually establish.
Displacement of testimony. If digital evidence is treated as superior, testimony that conflicts with it may be dismissed. But the person who was there may know things no screenshot could capture: how she felt, what she understood, what she was able to do. Privileging digital evidence can thereby reproduce a familiar injustice, in which a complainant’s first-person account counts for less than an artefact.
Uneven availability. Digital evidence exists in some cases and not others, depending on whether anyone recorded or shared anything. If fact-finders come to expect it, cases without it may be treated as weaker, even though most sexual offences leave no digital trace.
Testimony is evidence too
It is worth remembering that testimony is not merely a weaker form of evidence. Philosophers of testimony, from Thomas Reid to contemporary writers, have argued that we rationally rely on what others tell us for most of what we know. In legal settings, testimony is tested through cross-examination and comparison with other evidence. Its weaknesses are known and can be managed. The weaknesses of digital evidence are less familiar and, partly because of its apparent objectivity, less often scrutinised.
How should they be weighed?
Neither form of evidence should be given automatic priority. Each should be assessed for reliability, completeness and relevance to the question at issue. Digital evidence is often highly valuable for establishing what was communicated and when. Testimony is often essential for establishing what was meant, what was experienced and what happened outside the frame. The strongest findings typically come from convergence, where testimony and digital records support each other, and the most difficult from conflict, where neither can simply override the other.
In this case, the group chat may establish who received what and when, which is relevant to the responsibilities of bystanders. Whether it establishes what happened in the room, and whether there was consent, is a different matter, on which the testimony of those present will remain central.
Conclusion
Messages, photos and screenshots should not carry more weight than testimony simply because they are digital. They have distinctive strengths, especially contemporaneity, but also distinctive risks: fragmentation, ambiguity, manipulation, false precision and the displacement of first-person accounts. The epistemic task is not to rank kinds of evidence in advance but to scrutinise each for what it can and cannot show, and to resist the persuasive force of an image that seems to settle a question it does not actually answer.