Citation: R v Youseff (No 2) [2024] NSWSC 1260
Court: Supreme Court of New South Wales
Date: 8 October 2024
Judge(s): McNaughton J
Background
The accused stood trial for the murder of a man shot and killed at a residential address in Tregear in the early hours of 30 November 2020. The Crown's case was largely circumstantial, relying on evidence of a drug debt, alleged threats, phone tower data placing the accused's phone near the scene, and sightings of a car similar to his in the surrounding streets.
When police arrested the accused and searched his home in August 2022, they found two iPhones containing identical data in the Apple Health application for the period around the shooting. The Crown sought to tender a Cellebrite extraction report showing activity sensor readings, including steps taken, distance travelled, and flights climbed, between 2am and 3am on 30 November 2020.
The Crown sought to rely on this data without calling expert evidence about how the iPhone activity sensor works or how the data is generated. The defence objected well into the trial, prompting detailed consideration of the admissibility of this class of evidence.
Legal Issues
- Whether the Cellebrite extraction report containing iPhone activity sensor data was admissible under the Evidence Act 1995 (NSW), including under sections 146, 147, 48, 69, 144, and 183
- Whether the activity sensor step count data was relevant within the meaning of sections 55 and 56, given questions about its accuracy and temporal precision
- Whether, even if otherwise admissible, the evidence should be excluded under section 135(b) as evidence whose probative value is substantially outweighed by the danger of it being misleading
Decision
Issue 1: Statutory pathway for admissibility
McNaughton J observed that the Crown relied solely on section 146 of the Evidence Act, but noted the evidence did not sit easily within any single provision. His Honour assumed, for the purposes of analysis, that the digital record of the iPhone activity sensor data could be treated as a business record of Apple, making section 147 the more appropriate provision. Section 147 creates a presumption that a device or process produced a particular outcome, but his Honour found that presumption addresses only whether the device generated the recorded data, not whether that data accurately reflects real-world steps taken at the recorded times (at [22]-[32]).
Issue 2: Relevance and accuracy of the step count data
His Honour referred to a peer-reviewed 2019 article from the Netherlands Forensic Institute examining iPhone Health App data, which he located independently and provided to the parties. That article indicated that while average step count error was around 2%, the data was subject to significant variation and a risk of false positives, including from sources such as speed bumps (at [36]-[40]). Critically, the Cellebrite report recorded data at approximately 11-minute intervals, meaning any steps registered could have been taken at any point during the preceding 11 minutes, with no indication of the precise timing (at [41]). The court found that the data therefore could not be described as an accurate record of steps taken at a time proximate to when they actually occurred, and was not relevant for the purpose the Crown sought to use it (at [42]).
Issue 3: Section 135(b) exclusion
Even accepting, for the sake of argument, that the evidence was relevant and admissible, McNaughton J found it should be excluded under section 135(b) because its probative value was substantially outweighed by the danger of misleading the jury (at [43]-[44]). A data point showing 71 steps recorded at 2:35:15am, which was precisely when the shooting was alleged to have occurred, could in fact reflect movement up to 10 or 11 minutes earlier. Presenting that data without that qualification carried an obvious risk that the jury would treat it as confirming the accused's presence or activity at the moment of the shooting (at [44]).
Issue 4: Distance data
The Crown withdrew reliance on the distance travelled data after the peer-reviewed article was provided to the parties. That article demonstrated significant unreliability in the distance measurements (at [17]).
Orders Made
- The Cellebrite report purportedly depicting the activity sensor data recorded on an Apple iPhone is excluded.
Key Takeaways
-
iPhone Health App activity sensor data extracted via a Cellebrite report does not fit neatly within any single provision of the Evidence Act 1995 (NSW), and the Crown's reliance on section 146 alone was insufficient to secure admissibility in these circumstances (at [1], [22]-[23]).
-
Where digital device data is tendered for its accuracy, the presumptions in sections 146 and 147 go only to whether the device produced the recorded output, not to whether that output accurately reflects real-world events at the times indicated (at [32]).
-
Step count data recorded at approximately 11-minute intervals could not establish activity at a specific moment in time, rendering it irrelevant for the purpose of placing the accused at the scene at the moment of the shooting (at [41]-[42]).
-
Under section 135(b), evidence may be excluded where its probative value is substantially outweighed by the danger of misleading the fact-finder, and the 11-minute data window created precisely that danger in this case (at [44]).
-
The absence of expert evidence explaining how the iPhone activity sensor works, how it generates its data, and what the output means was a significant factor in the court's inability to assess reliability and relevance, particularly where the evidence was contested late in the trial (at [1], [33]).
Legislation and Cases Referenced
Legislation
- Evidence Act 1995 (NSW), ss 48, 55-56, 69, 135, 144, 146-147, 183
Cases
- Norvill v Stokes [2006] NSWLEC 622