Apple Secretly Censors Over 100,000 Words on iOS Keyboards, Investigation Reveals

A Reuters investigation has revealed that Apple maintains a secret, multilingual blacklist containing over 100,000 words and phrases that iOS keyboards deliberately refuse to suggest in predictive text or process through swipe typing. The discovery was brought to light by security researcher Derek Kaufman, founder of the scientific recognition startup Blacklight, and independently verified by international cybersecurity experts.

While iPhone users can still type any word letter by letter, the blacklisted terms never appear in predictive prompts or swipe gestures. According to Kaufman, while many blocked terms involve profanity, drug references, or historical slurs like "Auschwitz" and "Al-Qaeda," a significant portion ventures into sensitive political and ideological territory. Phrases such as "Dictator Trump, Sleepy Joe," and "Right wing conspiracy" are heavily suppressed or automatically altered during swipe inputs.

Political suppression extends globally, affecting regional sensitivities. Terms like "Tibetan independence" and "United Ireland" have been removed from predictive engines. In French, swipe typing restricts "excessive immigration"; in Hindi, it blocks insults targeting Prime Minister Narendra Modi and opposition leader Rahul Gandhi; and in Hebrew, the keyboard suppresses the phrase "occupied territories."

The issue stems from how predictive text and AI engines analyze swipe gestures. When a user swipes across the screen, Apple's algorithms attempt to guess the intended word. If the term is blacklisted, the system forces a substitution, automatically turning "United Ireland" into "United Iceland." Security expert Matt Swisch of Tolmo cross-referenced Kaufman's findings against system code, confirming a near-perfect match.

In response, Apple declined to comment directly on the list's scope or ideological inclusions, stating that users retain the ability to type any word manually and that the keyboard aims to avoid offensive suggestions. However, civil liberties groups have criticized the opacity of the mechanism, with Electronic Frontier Foundation senior staff attorney David Green calling the secret filtering a chilling form of subtle censorship that strips users of transparency and control.