Swamp devil

Social media may be a swamp, but it is also a Garden of Eden for evidence when it comes to civil and criminal disputes. Dr David Cowan hacks through the undergrowth

Social media is a terrain that needs careful navigation by solicitors and clients alike.

Adding to the swamp is the AI ‘slop’ that makes evidence produced through social media harder to navigate due to the volume, quality, and unreliability of the information shared.

Social media has become the digitally powered witness of crowds, and the smartphone the digitally powered personalised witness. In either case, the client cannot assume the digital witness is on their side.

Solicitors need to have a sophisticated social-media strategy specific to evidence if clients are to have protection from the dangers of the swamp.

Killing floor

Social media is an area where the hackneyed phrase ‘provides threats and opportunities’ is more than usually apt. The threat lies in how it is weaponised and in doubts as to the credibility of content shared – but the same goes for the opportunity it offers.

It is a classic double-edged sword, and it all turns on the credibility of content, which is increasingly harder to gauge.

How people use social media can both help and hinder lawyers in gathering and testing evidence. If the information is credible, then it can be used to help clients. If it is not credible, then it may also be used to help clients if the social-media content is shown to be false in respect to other parties.

If a client has abused social media, then it is all threat.

Social-media posts – whether words, images, or reactions – are malleable. The poster puts something online in a specific context with a particular meaning at the time. It is time that becomes the enemy, because things change and social media suffers from two defects in this respect.

First, even though an image post may be timestamped, it does not prove when the picture itself was taken. Second, an image posted in one context is notoriously capable of being decontextualised and thus can be easily weaponised.

This can make the content unreliable as evidence, on both sides. What started out as a spontaneous or casual post can become an exhibit in evidence. This means clients need to be careful of what they put online at all times. They need to be doubly vigilant during the time of a disputed matter.

Tagged photos, feeds from friends and family-account feeds, comments on news stories, or other posts on websites can all be fodder for opposing parties. They can come back to bite the poster on the ‘bahoochie’, as the Scots would say.

The poster cannot always rely on the use of the image or may find that, in the course of time, different interpretations become attached. This can be just as awkward from the other side of a dispute.

This is well illustrated in personal-injury cases. Insurance companies don’t particularly want to pay out, and legal teams can scour social media to find images of a customer enjoying themselves when they are supposed to have had a whiplash injury.

The difficulty can come down to proving when the picture was taken. A picture from more joyous times may be posted after an injury has occurred: the insurance legal team think they have a ‘gotcha’ moment, while the customer needs to set about proving it was an old photo.

In discovery, this can all be resolved but, in the meantime, billable hours have been spent and, perhaps, the damage done – after all a picture tells a thousand words.

How solicitors can advise clients on social-media use during a dispute (and before, if they are wise):

1. Audit: Identify all social-media accounts used and in use. Identify AI-generated material versus client-authored content.
2. Gather social-media content: Ask for as much material as possible to be provided, bearing in mind that what the client has deleted has not necessarily disappeared.
3. Content analysis: Review the content in light of the matter at hand, noting the context in which the material was generated, and any AI use.
4. Use analysis: Analyse the client use and online behaviour – get a picture of how the client used social media, regularity, and any monetisation of material.
5. Negative analysis: Discuss problematic material with the client and how it could be used or dealt with in terms of evidential value.

Do the do

Social media requires a specific case strategy that covers each account, platform visited, username and handle used. The strategy may require the client to ‘go dark’ during the dispute.

This does not mean that they are doing something wrong. Rather, it is a protection strategy to ensure they are not adding evidence to the matter. In other words, do not fuel the flames or feed the beast.

Going dark, or saying nothing, is a hard media strategy. If the client feels wronged, they will want to clear their name. However, I have far too often seen situations in media work where this instinct has led to disastrous outcomes.

Digital material needs to be preserved, and deleting content does not mean it has been effectively disposed of: it simply means that the client record has been deleted. The data can remain stubbornly on servers elsewhere. It may be difficult to retrieve by the client and opposing parties alike, but it is not impossible.

The obverse is to delete nothing. Although deletion can make material harder to trace, the fact that it does not disappear means that, when it comes to light – whether through active search or inadvertence – it may well have increased evidential power because it has been deleted. The deletion itself becomes interpretative.

Another aspect of social media is that its very nature means that there is no such thing as a private post or social-media space.

There are many and various circumstances through which any content may be made public, even if private accounts are used. Social media needs to be treated the way microphones have long been viewed – treat it as ‘live’.

I ain’t superstitious

The issue of spoliation has long applied to electronic evidence and applies equally to the digitally generated evidence of smartphones and social media.

Distinguishing client and adviser content from AI content can be helped by using AI watermarks.

In this way, it acts like a digital signature that singles out genuine content and verifies authorship. If such content is then repurposed via AI use, then this can be detected and set apart from what is the original content.

Liking and sharing become evidence to be interpreted. Is something shared implicitly approved, or is it a matter of mere interest or current mood?

An online search, either by the client or the solicitor, can create digital evidence. It leaves a trail. On LinkedIn, for instance, if you search for a party, then your search may be made transparent to the other party if you have not disabled that function in your search.

The problem that AI poses is also that it generates its own data that can be interpreted. Thus, content and records can reveal patterns of behaviour, in addition to storing direct statements that the client has made or recorded.

AI takes an individual’s smartphone use and online activity, generates its own reports, and uses this learned data to inform how they are using and what may be of interest to the user, based on their past usage.

In other words, the smartphone becomes sleuth and profiler.

This is about more than simply leaving a digital footprint. It leaves digital skeletons as well! The art of AI forensics is increasingly sophisticated, both in terms of the generation of AI forensic material created and the forensic tools used to track down social-media use.

I asked for water

It is important to keep records of advice related to social media provided to the client. At a later point, a client may ignore the advice, purposively or inadvertently, and then seek to blame the legal adviser.

Such records can protect both sides and demonstrate the seriousness with which both adviser and client are taking the social-media threat.

The essential advice should include handling of claim-related digital material, history and deletion management, and privacy settings.

Social media and AI are constantly changing, which means monitoring of all these matters needs to be undertaken on a regular and vigilant basis.

The biggest threats emerging include the use of deepfakes, cybersecurity concerns, and AI interpretation of data. This means having an effective and constantly evolving social-media policy. Such a policy should be built on four key principles that are of use to the practice and client alike:

1. Complaints handling: should be systemic, monitored, and acted upon.
2. Means of monitoring: track changes in regulation of illegal and harmful online content, which means historical tracking for threats of decontextualised content.
3. Protection measures: content, promotion, and advertising targeted at minors, vulnerable, and disadvantaged persons, including the ‘digital age of consent’ and other permissions.
4. Digital literacy: policy checked against provisions of the EU AI Act, Online Safety and Media Regulation Act 2022, Digital Services Act, Digital Markets Act, and Data Protection Act 2018. Increased AI literacy, interdisciplinary approaches, and avoidance of silos of information internally.

The social-media space and technology do not stand still, so policy cannot be static either.

Dr David Cowan is assistant professor in law at Maynooth University and the author of Law and Technology (Bloomsbury Professional Ireland, 2025) and Effective Communication for Lawyers: A Practical Guide (Elgar Law, 2022).

The do’s and don’ts

Do:

• Preserve everything.
• Check privacy settings.
• Record all advice given – for the mutual benefit of lawyer and client.
• Use AI watermarks.

Don’t:

• Delete anything.
• Assume searches are closed.
• Advise or agree to a client’s removal of social-media, claim-related material.
• Scrub, alter, or conceal a client’s digital history.
• Assume that privacy settings put content beyond discovery.