Publié : 29 September 2025
Actualisé : 1 month ago
Fiabilité : ✓ Sources vérifiées
Je mets à jour cet article dès que de nouvelles informations sont disponibles.
📋 Sommaire
📋 Sommaire
The technological landscape is buzzing with generative artificial intelligence, and Apple, despite its penchant for secrecy, is deeply involved. Recent revelations from Bloomberg, spearheaded by journalist Mark Gurman, have shed light on a fascinating initiative: the Cupertino giant is actively developing and testing an advanced chatbot application, designed to rival popular platforms like ChatGPT and Perplexity.
However, what proves most surprising is not merely the existence of this cutting-edge technology, but Apple’s peculiar decision to keep it out of reach for its millions of users. As the company gears up for a major overhaul of its Siri voice assistant by Spring 2025, this text-based conversational application, internally dubbed “Veritas ,” appears set to remain a closely guarded secret, sparking questions about Apple’s overarching AI strategy.
🔥 The Veritas Enigma: A Revolutionary Internal Chatbot
Reports indicate that an application named Veritas is currently pre-installed on the iPhones of select Apple employees. Its objective is clear: to offer a written conversational chatbot experience with impressive capabilities. Beyond simple text exchanges, Veritas reportedly allows users to revisit previous conversations, execute actions within other applications, edit photos, and even search through user data for contextual queries. This represents a natural language interface that could profoundly transform interaction with an iPhone.
The existence of Veritas demonstrates Apple’s technical prowess in creating a direct competitor to leaders in conversational AI. Nevertheless, Gurman suggests this application primarily serves as a crucial testing ground. It is reportedly used to evaluate the performance of a hybrid model, combining Apple’s internal LLM with an external model (potentially Google Gemini), as part of “Project Linwood.” Feedback from this internal deployment is vital for refining the future version of Siri.
🔥 The Promise of Siri 2.0 and Hybrid Models
For years, Apple users have eagerly awaited a version of Siri capable of competing with the rapid advancements in generative AI . Following delays, a significant revamp is anticipated for Spring 2025, although the newsletter mentions March 2026 for Project Linwood’s full maturity. This new Siri is envisioned as more than just a voice assistant; it aims to be an intelligent, proactive interface, leveraging a large language model for more nuanced and contextual interactions.
At the core of this transformation lies “Project Linwood,” focused on integrating a hybrid model. The concept is to capitalize on the local processing power of Apple devices for certain tasks, ensuring privacy and speed, while drawing upon the vast knowledge base of an external model, such as Google Gemini, for complex, up-to-date information requests. This approach promises a balance between performance and user privacy.
“Apple’s ambition is clear: not to create another generic chatbot, but to seamlessly merge the power of generative AI with the iconic iPhone user experience, transforming Siri into a truly intelligent and integrated assistant.”
🔥 A Puzzling Strategy Against Competition
Apple’s decision not to launch Veritas as a public application is particularly puzzling given that apps like ChatGPT and Gemini rank among the most downloaded on the App Store. Users clearly show an insatiable appetite for these conversational interfaces. By foregoing a dedicated application, Apple appears to overlook a clear market demand, preferring to channel AI access primarily through Siri, whether via voice or text.
Furthermore, the Cupertino firm has announced that from iOS 26 onwards, third-party developers will be able to utilize its own foundation model. Why offer this capability to partners while refusing to launch its own flagship application, especially if it benefits from a superior model as suggested by Veritas’s development? This approach could potentially empower competitors by allowing them to dominate the dedicated chatbot app space.
🔥 Implications and Future Outlook
Massive AI integration is anticipated as early as March 2026, or potentially with iOS 27. However, if Siri remains the sole primary access point for generative AI for the general public, Apple risks frustrating a segment of its users who seek pure, direct textual interaction with a chatbot. This could also weaken its standing against competing ecosystems that already offer robust, dedicated applications.
This strategy raises a critical debate: Is Apple aiming for such seamless integration that users won’t feel the need for a separate app, or is this a missed opportunity to capture a significant share of a booming market? The stakes are high, as the AI user experience will be crucial for retaining customers and enhancing the attractiveness of future iPhones.
📊 Summary of Apple’s AI Developments
Time will tell if Apple reconsiders its stance before the widespread deployment of AI across its systems. In the interim, the development of Veritas underscores Apple’s technological capability, but its internal containment raises questions and marks a notable divergence from current trends in the conversational AI market.
❔ Frequently Asked Questions
Why does Apple choose to keep “Veritas” internal and secret when a major Siri redesign is planned, and generative AI competition is fierce?
How does Veritas’s ability to “mine user data for contextual searches” align with Apple’s uncompromising reputation for privacy protection?
What are the technical and security implications of Veritas’s ability to “execute actions within other applications” and “manipulate photos” within the iOS ecosystem?
Does the fact that Veritas is an internal, non-public chatbot indicate a deeper, more native generative AI integration approach than its competitors, as opposed to a standalone application?
0 Comments