By: Tristan Jenkinson
The webinar contains many pearls of wisdom from Doug, Tom O’Connor, Director of Gulf Coast Legal Technology Center and Mike Quatararo, President of ACEDS. But there was one specific area that I wanted to follow up on because I had been discussing this with colleagues recently – the potential drawbacks of non-native disclosure, such as image only, compared to native disclosure.
One of the reasons this is of particular interest to me is that here in the UK the production rules are a little different. The standard in the Courts of England and Wales is to produce natively (or near natively where appropriate). Although non-native productions are still in use, they can ring alarm bells unless there is a specific reason for using them.
Breaches of Disclosure Obligations
Although thankfully rare, I have been approached to investigate a number of cases where parties have been suspected of not meeting their disclosure obligations. This includes cases where information that parties had originally tried to exclude has later been used to evidence breaches of court orders, falsification of material evidence, or large scale manipulation of data prior to disclosure. These investigations are into, allegedly, deliberate attempts to affect the outcome of a case. The murky depths of failures with data collections or issues with how searches are run is a topic for an entirely different blog.
One popular method that I have seen used to limit the information being provided, is to use image only productions. If you pair this with providing limited metadata in the load file, a lot of information can be conveniently excluded from the disclosure exercise.
Doug Austin and Craig Ball on Image Only Productions
The ACEDS webinar mentions two articles by Craig Ball from earlier this year, which I would highly recommend taking a look at.
Those posts stem from the Kamuda v Sterigenics case where the Judge (Christopher E Lawler) ordered that parties would produce their disclosures in native format, rather than in image format which was proposed by the defendants. The plaintiffs had argued that the format proposed by the defendants would be more expensive, imposing costs which would be unreasonable and act as a barrier to them making their case. Craig was retained by the plaintiffs to testify, supporting their position.
In his post (which you can find here) Craig explains that image only productions are typically more expensive than native productions. He goes on to discuss that the main reason for this is the filesize of a TIFF file compared to the native filesize, and the subsequent cost implications arising from that for processing and hosting. There is also often a charge for printing to TIFF.
Shortly after publishing the above article covering costs, Craig wrote a further article where he discussed how TIFF productions are often also less useful than native productions when it comes to searching. Craig covers a number of excellent points, in particular with regard to comments and track changes. You can read Craig’s article in full here.
Native Productions as Standard
As discussed above, in the Courts of England and Wales, the vast majority of our disclosures are in native format (with emails, for example, in near native format such as .MSG). This is driven by the Civil Procedure Rules which (in Practice Direction Part 31B, paragraph 33) state that;
“Save where otherwise agreed or ordered, electronic copies of disclosed documents should be provided in their Native Format, in a manner which preserves Metadata relating to the date of creation of each document.”
This means that here, non-native productions are far more unusual compared to other jurisdictions such as the United States.
Printing TIFF Images Can Reduce Usability
This is another point that Craig covers, with a particular focus on searching which I won’t duplicate here. There are however other drawbacks with printed TIFF files in a number of scenarios.
For example, if you have a large Excel spreadsheet, printing this out to TIFF images could lead to thousands of pages. This then requires an approach akin to putting together a jigsaw if you need to find some corresponding values across the file – compared to how easily you could identify the value you’re looking for within the corresponding native Excel file.
There are also other issues with printing Excel files, such as only the values being displayed, rather than the content of formulas. While this would likely not be an issue in the vast majority of cases, it could be key and does demonstrate a loss of usability compared with the original native files.
A Relevant Question (Especially in Construction Disputes) – Extracted Text or OCR Text?
As discussed above, in the UK, native files are the default for productions. There is one usual exception – redactions. It makes sense to take a momentary step back to discuss the approach to redactions in the UK.
For redacted documents (and potentially their families), the document typically has the redaction applied and is then printed as a TIFF or PDF file. This image is then produced, as opposed to providing the native file. This leaves an additional problem – what should be done with the extracted text?
If the reason for redaction was the textual content (which is most cases) then providing the extracted text for the document in the production renders the redaction process effectively redundant. For this reason it is common to perform an OCR (Optical Character Recognition) of the image of the non-redacted text, then supply the resulting OCR Text as the extracted text for the document.
It should be noted that while OCR is a great technology in making non-searchable files searchable, it is not (yet) perfect. The quality of the text it produces is dependent on many factors, including the quality of the original image, the language in use, if multiple languages or alphabets are in use and if there are multiple columns of text. Even in favourable situations, OCR may not be 100% accurate.
This then raises a relevant question – if you are receiving a TIFF production, rather than a native production, is the text supplied for each document the actual extracted text from the native or is it the result of performing OCR on the image of the native?
If the latter, then this could be a concern due to the limitations of OCR discussed above.
These problems could be compounded depending on the settings that have been used to create the TIFF images. With large documents such as CAD drawings (which are common in construction disputes) the resulting image could be shrunk to fit on a page. This can have a significant effect on the legibility of any text that may be present on the document which could then be lost, or become illegible once OCR is applied.
Additional Thoughts with CAD
While on the topic of CAD files, these can bring along their own problems when it comes to non-native productions. Often such files can be multi-layered, or a series of overlaying schematics within a single file. Depending on how the image is produced, you may receive just a single image – therefore losing much of the information in the separate layers.
This is a similar concept to that presented by Craig who describes this as;
“… challenges inherent to dealing with three-dimensional data using two-dimensional tools. Native applications deal with Comments, speaker notes and formulae three-dimensionally. We can reveal that data as needed, and it appears in exactly the way witnesses use it outside of litigation. But flattening native forms to static images and load files destroys that multidimensional capability.”
While not a consideration in the majority of electronic disclosure cases, there are cases where there are concerns or allegations of falsification of a document, email, or image in the disclosure process.
Where original native files are provided, these can provide additional artefacts that could potentially be used to investigate the authenticity (or otherwise!) of a file. Such investigations however will likely not be possible if the file is provided in TIFF form, as the underlying data of the file is not available to interrogate.
For example, if only TIFF versions of email files are being provided, it would be relatively simple to falsify the content, or, for example, change the date displayed on the TIFF file. If a native copy of the email were provided, then an expert may be able to examine the file to identify evidence to support or refute the authenticity of said email.
Similarly with photographs, if only TIFFs are provided, then potentially relevant EXIF information about when the photograph was taken, with which device etc. would be lost. Such information could, and has been, crucial in cases such as insurance fraud allegations.
Word documents and other Microsoft Office files can similarly be analysed for information which would not be available within TIFF copies of the file.
Therefore, where there are concerns about potential fraud, the suggestion of an image only production is often a red flag.
It’s worth noting that there is a case in the US where the above has been important. In the Kleiman v Wright case (which I will write more about separately but you can see the court documents here) an email was provided as a PDF. That PDF was analysed and evidence was identified that suggested that the PDF itself was manually altered to change the date on which it was sent (amongst other details). Had the PDF been subsequently printed to TIFF, the same analysis would not have been possible to complete.
In the UK, native productions are the standard. This means that requests for non-native productions such as image only productions can be seen as immediate red flags. Before agreeing to receive a non-native production, or making your arguments against, you should carefully consider what information you might be missing out on. Similarly, if you are suggesting making an image only production, consider carefully what your reasons are so that you can provide these should the other side resist.