Loading story…
Apple released iOS 27, macOS 27 Golden Gate, and companion updates for watchOS, visionOS, and tvOS, led by a context-aware Siri AI built on Apple’s AFM 3 models. The assistant can interpret natural-language requests, explain on-screen content, search personal context including emails and messages, and invoke supported apps; Apple also added AI-generated Shortcuts, Safari extensions, photo-editing features, and Liquid Glass interface controls. Private iOS and macOS frameworks reportedly show an architecture for delegating Siri requests to third-party models and potentially replacing Apple’s server-side model, although the currently available interface exposes only ChatGPT and Apple has not opened the relevant entitlement to external developers.
The releases also expand child-safety controls through Family Sharing Child Accounts, including Ask to Browse approvals for new websites, approved-contact requirements, category-based time allowances, scheduled app access, and Communication Safety detection for violent and graphic material. Apple’s iOS 27 privacy notice additionally offers users an opt-in to contribute Siri interaction audio and transcripts to improve Siri AI and train Apple language models, reversing earlier assurances that private interactions would not be used for foundation-model training. Organizations should assess Siri AI’s access to personal context, external-model integration exposure, and user privacy-consent settings before broad deployment.
Track how attackers are adapting to this technology.
7 events from the most recent confirmed update back to the earliest known activity.
Apple made redesigned child-safety controls available for iPhone, iPad, and Mac, requiring iOS 27, iPadOS 27, or macOS 27. The rollout added Ask to Browse, approved-contact requirements, category-based Time Allowances, scheduled app availability, and expanded Communication Safety protections for violent and graphic content.
Apple released iOS 27, macOS 27 Golden Gate, watchOS 27, visionOS 27, and tvOS 27. The releases added the context-aware Siri AI overhaul on all listed platforms except tvOS, alongside Apple Intelligence capabilities powered primarily by AFM 3 models.
Code researcher pdfu identified private iOS 27 and macOS Golden Gate frameworks for Model Delegation and external inference providers, indicating Siri could delegate tasks to third-party AI models. The currently shipping macOS release candidate exposed the Ask interface only for ChatGPT, with no third-party Model Delegation entitlement available.
An iOS 27 release-candidate privacy notice said Apple may collect opted-in Siri and dictation audio, transcripts, requests, and responses to improve Siri AI and train its in-house language models. Apple revised its responsible-AI language to allow use of private interactions when users explicitly opt in.
Apple previewed an overhaul of parental and child-safety controls for its devices, centered on Child Accounts managed through Family Sharing.
Apple later reinstated human evaluation of Siri recordings under an opt-in program following its earlier pause.
Apple temporarily stopped human evaluation of Siri audio recordings after reports that contractors had reviewed sensitive and unintentionally captured conversations.
Vulnerabilities, threat actors, malware, products, organizations, and breaches Mallory has linked to this story.
Follow how adversaries are adapting to this technology, and where it touches your stack today.
4 references tracked. Mallory keeps watching after this page renders.
helpnetsecurity.com
Open sourcearstechnica.com
Open sourcemacrumors.com
Open sourceheise.de
Open sourceMap indicators from this story to your assets and identify affected systems in minutes.
Every observed campaign, victim, and pivot linked to actors named in this story.
Malware, exploits, and IOCs connected to the activity described here.
YARA, Sigma, and Snort rules deployed to your SIEM as soon as they’re published.
Get matching new stories delivered to your team as they break — not the next morning.
Ask questions about this story and take action on the answers.
Other stories Mallory is tracking that overlap on entities, sources, or topic labels.
Apple executives used WWDC and a subsequent interview to position the company’s AI strategy around practical assistance, privacy, and tight product integration rather than emotionally driven chatbot experiences. Craig Federighi said Siri is designed to help users complete tasks and retrieve information, while Greg Joswiak said Apple is not building AI for its own sake. Apple also highlighted privacy-preserving personalization through on-device processing and **Private Cloud Compute**, alongside security-focused features such as automatic breached-password remediation in the Passwords app and Safari and new child safety controls for highly restricted iPhones. At the same time, reverse engineers reportedly extracted and published Apple’s new AI-driven Siri system prompt from the iOS 27 developer beta, exposing more than 1,300 lines of instructions governing alignment, reasoning, multimodal behavior, and tool use. The leaked materials indicate Siri is designed to disclose limitations, resist adversarial prompt manipulation, avoid anthropomorphic behavior, and rely on structured JSON entities derived from device data such as contacts, email, calendars, and web searches, with actions constrained through deterministic function calls and clarification workflows. The publication of the prompt on public platforms has given AI engineers and security researchers an unusual look at how Apple is implementing the privacy- and safety-focused approach it is publicly promoting.
Jun 29, 2026
Apple released **iOS 26.4** and **iPadOS 26.4**, making **Stolen Device Protection** enabled by default and shipping **34 security fixes** across components including **Audio**, **iCloud**, **Mail**, **Printing**, **Siri**, and the **iOS kernel**. The update now requires biometric authentication to view saved passwords and disable Lost Mode on an iPhone, strengthening protections against device theft and account takeover. Apple also issued corresponding `26.4` updates for **watchOS**, **macOS**, **tvOS**, and **visionOS**. Alongside the security changes, Apple added several user-facing features across its platforms, including AI-generated **Playlist Playground** playlists in Apple Music, an **Ambient Music** widget, offline song recognition in Control Center, native **HLS** video episode support in Podcasts, average bedtime tracking in Health, and support for third-party AI chatbots in CarPlay once providers update their apps. The release also fixes a long-running keyboard accuracy bug, while the widely rumored upgraded Siri was not included in this version.
Jun 29, 2026
Apple announced new *age-assurance* capabilities aimed at complying with expanding child-safety regulations, centered on an updated **Declare Age Range API** that returns an age bracket (e.g., under 13, 16–17) rather than a precise birthdate. In certain jurisdictions (including Australia, Brazil, and Singapore), Apple also plans to **block downloads of 18+ apps** until the user confirms they are an adult via an App Store-managed flow, alongside OS-level family settings intended to enforce age-appropriate restrictions without pushing sensitive identity collection to individual apps and websites. Separately, Apple’s **iOS 26** and **iPadOS 26** (with specific hardware) were reported as receiving **NATO Restricted** approval, enabling use for classified information up to the Restricted level under an “Indigo” configuration listed in NATO’s Information Assurance Product Catalogue. The approval was attributed to platform security features such as encryption, **Face ID**, and **Memory Integrity Enforcement**, with additional emphasis on Secure Enclave capabilities in newer chips; the evaluation reportedly involved Germany’s BSI. A third item provides general guidance on **CUI enclaves** and **NIST SP 800-171** controls for protecting Controlled Unclassified Information on mobile/remote-access workflows, but it does not describe the Apple/NATO certification or Apple’s age-verification tooling as a specific event.
Jun 29, 2026