As a victim or suspect of a crime, or witness to an offence, you may find your actions, behaviour and character scrutinised by the police or a barrister using CCTV footage. You may assume all the relevant footage has been gathered and viewed. You may sit on a jury and be expected to evaluate CCTV footage to help determine whether you find a defendant guilty or innocent.
You may believe you will see all the key images. You may trust the camera never lies.
However, the evidence we gathered during our study of British murder investigations and trials reveals how, like other forms of evidence such as DNA and fingerprints, CCTV footage requires careful interpretation and evaluation and can be misleading.
Instead of providing an absolute “truth”, different meanings can be obtained from the same footage. But understanding the challenges and risks associated with CCTV footage is vital in a fair and transparent system to prevent possible miscarriages of justice.
The justice system often relies upon digital evidence to support investigations and prosecutions and CCTV is one of the most relied upon forms. Recent estimates suggest there are more than 7.3 million cameras in the UK, which can capture a person up to 70 times per day.
The public may be filmed on council-owned CCTV, by cameras in commercial premises, or at residential premises (home cameras or smart doorbells, as well as on public transport and by dash cams.
In our study of 44 British murder investigations, we showed how CCTV provides many benefits to investigators. It can help identify suspects and witnesses, and implicate or eliminate suspects. It can also help to corroborate or refute accounts provided by suspects and witnesses. However, our findings also indicate how CCTV can be unreliable and problematic.
CCTV is sometimes inaccessible or lost because the detective who is sent to retrieve the footage lacks the skills, training or equipment to recover it in a timely manner. This is especially important since CCTV is often deleted within three weeks of being recorded. We found that it was often over-written within 7 to 10 days.
At other times, owners are unable to access systems or cannot manage the volume of CCTV requested, for instance, when taking buses out of service for footage to be downloaded. And even when footage is successfully seized, there may not be officers available to view it all.
There is also the risk that important footage which could exonerate a suspect is not disclosed to the defence, which could mean innocent people are imprisoned.
Detectives must frequently make sense of poor-quality images that are blurry or grainy. This is not easy. In some of the investigations we observed, the police tried to enhance poor-quality images, though this was not always successful.
Investigators must also decide whether to draw on experts to interpret footage and present evidence at court. However, the police have no clear guidance to help determine whether and when to draw on such expertise. We observed cases where officers decided against expert input because they were confident of their own interpretations.
Our study also revealed how some detectives or CCTV officers are used repeatedly to view or interpret footage because they are regarded by others (or assign themselves) as “super-recognisers”. These are people who may be better at recognising faces than others. However, there is no robust measure for determining whether someone is a super-recogniser. Furthermore, if super-recognisers are incorrectly viewed as expert witnesses, their evidence could be overvalued during a police investigation or at court.
By the time CCTV footage is shown to a jury, it has been choreographed carefully by the police and prosecution barrister. They are often adept at selecting, organising and editing footage into slick packages.
These techniques are also used by the defence who deliberate over whether to use moving footage or still images, at what speed to show the clips and at what point to add commentary. This is to demonstrate an “alternative truth” and provide a contested interpretation of the same footage. It might be difficult for juries to determine how the footage has been edited.
Murder investigations are generally regarded to be the gold standard of criminal investigation, due to the investment of time, resources and expertise. Nevertheless, we uncovered many challenges, errors and risks involved in the use of CCTV. These are likely to be even greater in other kinds of criminal investigation, where staffing and knowledge of digital evidence may be more limited.
The complexities of CCTV evidence need to be understood by everyone involved in handling, interpreting and presenting footage, as well as by those of us whose actions and accounts may be scrutinised on the basis of CCTV footage.
The challenges and risks identified here are likely to intensify as digital technologies advance – demonstrated by recent concerns with automated facial recognition technologies and the risk of deepfake videos.
This research was funded by The Leverhulme Trust.
This research was funded by The Leverhulme Trust.