Admissibility as evidence of photographs taken by a mobile phone

Question

It would appear that Samsung has admitted that images captured using their mobile phones' "Scene Optimizer" feature (which can be disabled) are not necessarily an accurate photograph of what the user thought they were recording.

https://www.theverge.com/2023/3/15/23641069/samsung-fake-moon-controversy-english-language-blog-post

Specifically, the saved image, after processing by the "Scene Optimizer", might be a "close equivalent" to what the user was looking at, but have originated at a different location on a different date. Or alternatively, it might be a composite incorporating indeterminate proportions of multiple original images.

As a general question, have any legal systems started to take seriously the possibility that a stored image might not be an accurate representation of the scene?

Answer 1

You are asking about what is known as "authentication" of real evidence.

Photographic and digital image evidence is not presumptively inadmissible, but nor is it automatically admissible. Rather, this evidence is conditionally admissible. A threshold to admissibility of an image is its authentication. This depends on "(1) their accuracy in truly representing the facts; (2) their fairness and absence of any intention to mislead; and (3) their verification on oath by a person capable of doing so" (R. v. Andalib-Goortani, 2014 ONSC 4690, citing a long line of cases adopting this standard in Canada, but I understand this approach to be the norm across the common law).

This requires that the party seeking to have the evidence admitted provide some evidence that the image is a fair representation of the facts portrayed. This would normally be via testimonial evidence from a witness (often, but not necessarily, the photographer).

On direct examination, the party seeking to introduce the photo as an accurate representation would ask the witness about how the photo was captured, whether there was any special image processing applied during or after capture, etc.

The camera's automatic process can be a viable avenue for cross-examination of the witness. If relevant, this could include questions about color correction, zoom settings, chromatic aberration, features such as "Scene Optimizer," etc.

Either party might also call other witnesses to explain how any automatic features work, the extent to which they might modify a scene, and whether the user would have had any indication that it was operating. They might also explain how to interpret any associated metadata that might indicate that special processing was applied.

It is not necessary that complete accuracy be established. Even non-digital cameras merely produce projections of a scene that might be warped or otherwise altered by the lens. What will matter is the nature of any modifications and whether the image provided to the court is substantially accurate with respect to the material details. As one court has said:

complete accuracy of a photograph or video is not a prerequisite to admissibility. So long as, despite a change or alteration, the image is substantially accurate, is a fair representation of what it purports to show and is not misleading, the image may be admitted in the discretion of the judge, upon the application of a probative value/prejudicial effect analysis.

Even if admitted, the weight given to the information in the image is ultimately a question for the jury or judge (when sitting alone as fact-finder).

Answer 2

It has long been known that digital data can be (somewhat) easily manipulated, you just have to know how. Nevertheless, digital data remains admissible in court in both criminal and civil proceedings. Any data can be attacked as unreliable, and in some cases a type of data might be inadmissible if it is completely unreliable.

It is up to the jury to decide whether they believe the particular piece of evidence, which means it is up to the attorneys to advance the best arguments possible that the particular item is / is not reliable. This is a question of very specific facts, so the courts have not generally rules that voice recordings or photographs are inadmissible because they might have been manipulated. The jury would have to look at the expert testimony to decide if the object was manipulated.

Answer 3

In the US, from precedent, it's admission can be successfully challenged by opposing council, if fine details are important.

In the recent and high-profile Rittenhouse trial, the defense objected when the prosecutor tried to zoom in on a video on his Ipad, because Apples proprietary and unknown zoom algorithm may introduce extra pixels and that was sustained. Later, the prosecution had to get an expert witness to explain exactly what zoom algorithm was used to enlarge a photo and how it didn't alter pixels, before it could be admitted: https://journals.library.columbia.edu/index.php/stlr/blog/view/417

Answer 4

The starting point is that photos are generally admissible (provided that the person who took them testifies so), regardless of what technology was used to take them.

It is up to the opposing party to suggest that the photos may be inaccurate or tampered with. Usually the person who took them will be cross-examined about what device they used and with what settings. The party may also call an expert witness to testify that the device/settings is known to produce inaccurate/modified picture.

The jury will then decide whether the picture is accurate enough or not.