Apple Faces Intense Backlash Over Flawed CSAM Photo Scanning Algorithm
Security researchers highlight dangerous flaws in Apple's NeuralHash algorithm after a developer reverse-engineers the code. Critics warn the on-device scanning system is vulnerable to hash collisions and potential government abuse.
Apple faces massive criticism over its newly announced NeuralHash technology, a system designed to scan user devices for child sexual abuse material. The algorithm identifies known illegal images by matching unique digital hashes provided by child protection organizations, triggering a manual review if an account contains 30 or more matching photos. Apple insists this on-device approach is more privacy-friendly than server-side scanning, claiming the odds of a false positive are just one in one trillion.
However, security experts and privacy advocates express serious concerns that highly resourced actors could abuse this system to implicate innocent people or detect other objectionable materials for authoritarian governments. These fears intensify after a developer successfully reverse-engineers the NeuralHash algorithm from iOS 14.3 and publishes the code online for anyone to test.
The public release of the code quickly leads to the discovery of a hash collision, a critical flaw where two completely different images produce the exact same hash. This vulnerability directly challenges Apple's assertions of accuracy and underscores the real risk of innocent users being falsely flagged to law enforcement before the technology even officially launches.