Public
- Public
- Network
- Groups
- Featured
- Popular
- People

Conversation

Notices

Embed this notice
David Nathanson (d_j_nathanson@mastodon.social)'s status on Wednesday, 21-Feb-2024 09:19:38 JST David Nathanson

Important decision from MA Supreme Judicial Court regarding "black box" algorithmic digital forensics. The prosecution tried to use iPhone "frequent location history" evidence against the defendant to prove his presence at the scene. The SJC excluded the evidence as unreliable because the prosecution could not prove the methodology was reliable or testable.
If prosecutors want to use algorithmic evidence against defendants, they should have to disclose source code. #Law
https://www.mass.gov/files/documents/2024/02/20/m13499.pdf
In conversation about 9 months ago from mastodon.social permalink
Attachments
1. A page from the opinion that reads: The Commonwealth also argues that because the FLH algorithm uses as its inputs location data sources generally regarded as reliable (such as GPS, CSLI, and Wi-Fi location data), the FLH data output by the algorithm consequently is also generally regarded as reliable. This argument misunderstands the Commonwealth's burden. "[W]hen proposed expert testimony uses a new theory, or new methodology to apply an accepted theory, the proponent must establish its reliability using a Daubert-Lanigan analysis." Davis, 487 Mass. at 455. See Commonwealth v. Camblin, 471 Mass. 639, 650 (2015), S.C., 478 Mass. 469 (2017) (although breathalyzer technology is generally accepted as reliable, where new type of breath test machine had not previously been tested and no court had considered the reliability of its source code, "the judge should have held a hearing to determine whether the source code and other challenged features . . . functioned in a manner that reliably produced accurate breath test results"). And I highlighted this portion: "In other words, even if the inputs used by the FLH algorithm are generally deemed reliable, the FLH data outputs are not ipso facto reliable, especially where there is not scientific literature or adequate testing to support reliability."
  https://files.mastodon.social/media_attachments/files/111/966/217/009/141/462/original/c6e88c5221eceb76.png
2. Untitled attachment

Public

Conversation

Notices

Feeds