Apple Begins Scanning Public iCloud Links for Harmful Content

Apple announced Thursday it would begin checking public iCloud links for harmful content starting with the next major releases—iOS 27.3, iPadOS 27.3, macOS 27.3 and visionOS 27.3. The change specifically targets scenarios where users share iCloud content such as photos in iCloud Photos or files in iCloud Drive by generating a link accessible to “Anyone with the link.” When users choose this option, Apple’s new system computes a cryptographic hash value of each image or file, compares it against a database of known “harmful” hashes, and blocks access if a match occurs. This approach avoids inspecting the contents themselves, addressing privacy concerns while enforcing moderation. ¹

The policy implements Section 67(a)(c) of California’s Age‑Appropriate Design Code Act, passed into law last November after extensive debate over balancing free speech considerations versus child safety online. The act obligates technology companies to moderate user‑generated content that could expose minors to inappropriate material. Apple’s move follows prior controversies involving Facebook’s Messenger Kids app, whose content‑moderation practices drew scrutiny back then due to alleged lack of sufficient protections regarding harmful material exposure. ¹

Apple has not yet disclosed how large its “harmful” hash database is or what categories of content it covers beyond violent or sexual imagery. Experts note that hashing allows efficient detection without reading actual data, thereby preserving user privacy on the server side. Malicious actors may attempt evasion techniques such as subtly altering images—adding imperceptible noise—to change the hash while keeping semantic meaning intact. Effectiveness therefore hinges entirely upon the accuracy and completeness of the database; any bias or false positives would have serious consequences for users who share innocuous media. Experts also caution that hashing cannot detect new or unknown forms of harmful content unless they are added to the reference set, potentially leaving a window for novel abuses.

For consumers, the change means that if an iCloud link contains a photo or file flagged in the database, the recipient will see a warning or be blocked from viewing the content altogether. Users can still opt out by disabling public sharing, but the default setting now includes automatic scanning. The update aligns with broader industry trends where platforms increasingly employ automated moderation tools to curb extremist propaganda, child exploitation, and other illegal content.

In practice, this could reduce the spread of disallowed images shared via social links while preserving user anonymity—since only hash values travel between servers. Yet the system may also generate false positives if legitimate media inadvertently matches a harmful hash due to accidental similarity or hashing collision. If such errors occur at scale, users might experience unwarranted censorship, raising questions about transparency and appeal mechanisms.

Overall, the move represents a significant step toward systematic moderation of publicly shared iCloud content, balancing user privacy against the need to protect vulnerable audiences from potentially illegal or exploitative imagery. The effectiveness of the approach will depend largely on the comprehensiveness of the database and the robustness of the hashing algorithm in resisting evasion attempts by malicious actors. Apple’s initiative underscores that large platforms are increasingly adopting automated detection tools under legal frameworks designed to safeguard children while respecting free expression.