United States v. Williams

83 F.4th 994

Opinion

Case: 22-10316 Document: 00516930184 Page: 1 Date Filed: 10/13/2023

United States Court of Appeals for the Fifth Circuit United States Court of Appeals Fifth Circuit

____________ FILED October 13, 2023 No. 22-10316 Lyle W. Cayce ____________ Clerk

United States of America,

Plaintiff—Appellee,

versus

Howard Sanford Williams,

Defendant—Appellant. ______________________________

Appeal from the United States District Court for the Northern District of Texas USDC No. 3:20-CR-638-1 ______________________________

Before Smith, Southwick, and Higginson, Circuit Judges. Jerry E. Smith, Circuit Judge: Howard Williams was convicted of sex trafficking of a child in viola- tion of 18 U.S.C. § 1591(a)(1) and (b)(2). On appeal, he challenges the intro- duction of evidence pulled from his cellphone using Cellebrite technology. He claims that the district court erred by permitting a police investigator to introduce the Cellebrite extract without first being qualified as an expert under Federal Rule of Evidence 702. On this matter of first impression, we disagree. When law enforce- ment uses Cellebrite to pull information from a phone and a lay juror would Case: 22-10316 Document: 00516930184 Page: 2 Date Filed: 10/13/2023

No. 22-10316

require no additional interpretation to understand that information, the party does not need to introduce the evidence through an expert. We affirm.

I. Williams, 37, met his 16-year-old victim, Jane Doe, on Tagged, a social media/dating application. She was ranting about her difficulties at home, so Williams invited her to go out and smoke marihuana, presumably to forget about her troubles. He picked her up at a school near her house and drove to the nearby Lake Arlington. While they smoked, she complained about her toxic home life and said she did not want to return home. He offered to get her a hotel room, and Doe agreed. Williams checked Doe into an extended stay hotel, and, for the next five days, Williams kept her high and drunk, had sex with her, and pimped her out using several websites. After five days, Doe left the hotel, went to the hospital with her mother, and reported that she had been sex-trafficked. A police officer inter- viewed her, and she provided a photograph of Williams from her phone. Police arrested Williams a few days later, and he gave a recorded statement. In that statement, Williams admitted that he bought the hotel room for Doe, had sex with her twice, “lined up dudes” for her, gave her drugs, collected money from her prostitution, and communicated with the customers. After his arrest, the police used a Cellebrite device to copy the infor- mation off Williams’s and Doe’s mobile phones. To use the device, an investigator merely plugged each phone into it and ran the program. The program pulled out the user data—including any messages, videos, or emails sent, received, or recently deleted—along with the apps used on the phone, and provided it to the police in an accessible, easily-navigable, and readable format. Williams’s Alcatel phone, being less technologically cohesive than a Samsung or Apple phone, required an additional file system extraction to copy everything on the device. But, beyond that step, the process and the

2 Case: 22-10316 Document: 00516930184 Page: 3 Date Filed: 10/13/2023

product were the same for the two phones. At trial, the prosecution introduced evidence taken from the Celle- brite extractions of both Williams’s and his victim’s phones, the recorded statement Williams gave to the police, testimony from Doe, messages obtained from another messaging service not at issue, and more. In a denied motion in limine, Williams objected to the use of the Cellebrite testimony without an expert witness to introduce it, and he renewed his objection at trial. On voir dire, the police investigator disavowed any particular knowledge of Cellebrite’s technology or of any malware that may have affected the data extraction. The district court then overruled Williams’s objection, and the investigator testified to (1) his certifications as a Cellebrite Operator and a Cellebrite Physical Analyzer, (2) the data-extraction process, and (3) the evidence he obtained. On appeal, Williams claims that it was reversible error for the court to admit the Cellebrite testimony without an expert witness and a finding of reliability. We disagree.

II. We review preserved challenges to evidentiary rulings for an abuse of discretion. United States v. Caldwell, 586 F.3d 338, 341 (5th Cir. 2009). “A trial court abuses its discretion when it bases its decision on an erroneous view of the law or a clearly erroneous assessment of the evidence.” Id. Further, an evidentiary error is harmless if it would not have a substantial impact on the jury’s verdict. United States v. Williams, 957 F.2d 1238, 1244 (5th Cir. 1992) (citing United States v. Bernal, 814 F.2d 175, 184 (5th Cir. 1987)). “Unless there is a reasonable possibility that the improperly admitted evidence contributed to the conviction, reversal is not required.” United States v. Okulaja, 21 F.4th 338, 344 (5th Cir. 2021) (internal quotation marks and citation omitted).

3 Case: 22-10316 Document: 00516930184 Page: 4 Date Filed: 10/13/2023

Federal Rule of Evidence 602 permits a witness to provide fact testi- mony if he has personal knowledge of the matter. Accordingly, lay witnesses can testify “so long as the witness does not base his or her opinion for scientific, technical or other specialized knowledge within the scope of Rule 702.” Caldwell, 586 F.3d at 348. To that end, in Caldwell, we permitted a software company employee to testify how a particular software operated without being qualified as an expert witness under Rule 702 because of the “prevalence of computer technology” and because the opinion the employee was providing was “within the realm of knowledge of the average lay person.” Id.

III. We find no error, much less an abuse of discretion. Williams claims that Cellebrite is a complex technology, ergo, the operation of Cellebrite requires specialized knowledge, and the introduction of a Cellebrite report demands qualification of a witness as an expert. But this ignores the basic realities of life. All the officer did was run a computer program. He offered no technical understanding of the machine or software; he did not write the program; and he did not opine on any application of specialized knowledge. During trial, the investigator explained that “[a]s an operator, I purely operate the machine. I am not privy to the programming or how it extracts data.” Thus, he explicitly disclaimed that he was offering expert testimony. This is the antithesis of Rule 702’s requirement of “scientific, technical, or other specialized knowledge.” Fed. R. Evid. 702(a). Rather, the investi- gator knew no more than anyone else who runs a program on his computer that he did not write. Every circuit that has addressed this question—whether evidence obtained with Cellebrite technology requires expert testimony for admission—has answered it in the negative. In Chavez-Lopez, the Fourth

4 Case: 22-10316 Document: 00516930184 Page: 5 Date Filed: 10/13/2023

Circuit concluded that the sponsoring witness offered only fact testimony, such as “the actions he took to extract the data—hooking the phones up to a computer, following a few prompts, and saving data onto an external drive.” 767 F. App’x 431, 434 (4th Cir. 2019). “At most, [the witness] offered the opinion that Cellebrite copies data from a cellphone, which he derived from his personal experience using the software.” Id. That testimony “didn’t require a technical understanding of Cellebrite, and he made no claims about the program’s effectiveness or reliability.” Id. Similarly, in United States v. Marsh, the witness merely described his training, explained how he used the software, and “confirmed the results by checking the [text] messages on the phone itself.” 568 F. App’x 15, 17 (2d Cir. 2014). And in United States v. Ovies, the Ninth Circuit reached the same result. 783 F. App’x 704, 707 (9th Cir. 2019), cert. denied, 140 S. Ct. 820 (2020). Reasoning similarly to the Fourth Circuit in Chavez-Lopez, the Ninth Circuit found that the witness’s testimony was not based on technical or specialized knowledge, as he “testified only about the steps he took using the Cellebrite program” and “did not opine as to the reliability” of Cellebrite. Id. An unpublished opinion of this circuit has also described a Cellebrite extraction as a “rote process” in a Confrontation Clause challenge. United States v. Rubio, No. 21-50886, 2022 WL 17246937, at *2 (5th Cir. Nov. 28, 2022) (unpublished). Williams claims that the Sixth Circuit reached a contrary result in United States v. Ganier, 468 F.3d 920 (6th Cir. 2006). But that misreads Ganier. First, Ganier does not mention, much less consider, Cellebrite. Sec- ond, even if it had, the computer reports at issue in Ganier are entirely dis- similar to a Cellebrite report. If, as in Ganier, we compare the Cellebrite report to other areas of expert testimony, this report is more like the “outputs of popular software programs” that an “average layperson today may be able to interpret . . . as easily as he or she interprets everyday vernacular” than the

5 Case: 22-10316 Document: 00516930184 Page: 6 Date Filed: 10/13/2023

“slang and code words used by drug dealers” a police officer needs “special- ized knowledge” to interpret. Id. at 926. We join our sister circuits. All the investigator testified to was how he downloaded the information from the phones using Cellebrite technology. At no point did he speak to the reliability of the software, except that he double-checked some of the report by looking directly at the source material in the phones themselves. To that end, he did not state any information on how Cellebrite operated in a technical sense, nor information that was beyond the knowledge of an average cell phone user. The investigator did not “impli- citly vouch for the accuracy or reliability of Cellebrite’s software,” as Wil- liams claims. Rather, the investigator acknowledged his lack of knowledge about the software and stated that he was merely an operator. Notably, the cross-examination of the investigator did not probe the reliability of Celle- brite technology. Without a showing of specialized knowledge, the mere use and under- standing of a Cellebrite extract at trial is insufficient to require an expert.1 Operating a Cellebrite device and understanding its report require knowledge in the realm of a reasonably tech-savvy lay person, regardless of the investi- gator’s testimony that he was a “certified” operator and analyzer. If we took Williams’s argument to its logical end, district courts would need a mechanic to qualify as an expert just to testify that a car with a flat tire ran a red light. That is not the requirement of Rule 702. The district court

_____________________ 1 Again, we highlight that the district court allowed voir dire of the investigator to address whether specialized knowledge would be imparted. Cf. United States v. Wehrle, 985 F.3d 549, 559 (7th Cir. 2021) (St. Eve, J., concurring) (“There are cases in which the collection of digital data would require testimony due, for example, to the sophisticated nature of the particular forensic analysis or the equipment deployed, or the technical nature of the testimony. But this is not one of those cases.”)

6 Case: 22-10316 Document: 00516930184 Page: 7 Date Filed: 10/13/2023

properly ruled that Rule 701 governs here.

IV. Williams also avers that the investigator’s testimony about his Celle- brite certifications indicated to the jury that the investigator was more trust- worthy than a typical lay witness. Admitting the investigator’s testimony was neither error nor an abuse of discretion. Even if it were, though, it was harm- less error. Evidentiary decisions are subject to harmless-error review, and “[u]nless there is a reasonable possibility that the improperly admitted evi- dence contributed to the conviction, reversal is not required.” Okulaja, 21 F.4th at 344 (internal quotation marks and citation omitted). The evi- dence of Williams’s guilt was overwhelming, particularly considering Willi- ams’s recorded confession. Thus, any erroneously admitted testimony by the investigator had, at best, a “very slight effect[,]” making it harmless. Kot- teakos v. United States, 328 U.S. 750, 764–65 (1946); see also Williams, 957 F.2d at 1244. AFFIRMED.