Respected university researchers are sounding the alarm bells over the technology behind Apple's plans to scan iPhone users' photo libraries for CSAM, or child sexual abuse material, calling the technology "dangerous."
Jonanath Mayer, an assistant professor of computer science and public affairs at Princeton University, as well as Anunay Kulshrestha, a researcher at Princeton University Center for Information Technology Policy, both penned an op-ed for The Washington Post, outlining their experiences with building image detection technology.
The researchers started a project two years ago to identity CSAM in end-to-end encrypted online services. The researchers note that given their field, they "know the value of end-to-end encryption, which protects data from third-party access." That concern, they say, is what horrifies them over CSAM "proliferating on encrypted platforms."
Mayer and Kulshrestha said they wanted to find a middle ground for the situation: build a system that online platforms could use to find CSAM and protect end-to-end encryption. The researchers note that experts in the field doubted the prospect of such a system, but they did manage to build it and in the process noticed a significant problem.
We sought to explore a possible middle ground, where online services could identify harmful content while otherwise preserving end-to-end encryption. The concept was straightforward: If someone shared material that matched a database of known harmful content, the service would be alerted. If a person shared innocent content, the service would learn nothing. People couldn't read the database or learn whether content matched, since that information could reveal law enforcement methods and help criminals evade detection.
Knowledgeable observers argued a system like ours was far from feasible. After many false starts, we built a working prototype. But we encountered a glaring problem.
Since Apple's announcement of the feature, the company has been bombarded with concerns that the system behind detecting CSAM could be used to detect other forms of photos at the request of oppressive governments. Apple has strongly pushed back against such a possibility, saying it will refuse any such request from governments.
Nonetheless, concerns over the future implications of the technology being used for CSAM detection are widespread. Mayer and Kulshrestha said that their concerns over how governments could use the system to detect content other than CSAM had them "disturbed."
A foreign government could, for example, compel a service to out people sharing disfavored political speech. That's no hypothetical: WeChat, the popular Chinese messaging app, already uses content matching to identify dissident material. India enacted rules this year that could require pre-screening content critical of government policy. Russia recently fined Google, Facebook and Twitter for not removing pro-democracy protest materials.
We spotted other shortcomings. The content-matching process could have false positives, and malicious users could game the system to subject innocent users to scrutiny.
We were so disturbed that we took a step we hadn't seen before in computer science literature: We warned against our own system design, urging further research on how to mitigate the serious downsides....
Apple has continued to address user concerns over its plans, publishing additional documents and an FAQ page. Apple continues to believe that its CSAM detection system, which will occur on a user's device, aligns with its long-standing privacy values.
