Apple may run the next Siri on a custom Google Gemini model
Recent reports claim Apple has asked Google to build a specially adapted version of its Gemini AI to power the next major Siri update. If confirmed, the move would represent an unusual partnership: Apple relying on a Google‑developed model for a core assistant feature rather than using only Apple‑built AI.
Details are sparse. Key unknowns include where the model will run (on‑device vs. cloud), whether Apple or Google will host inference, and how Apple plans to handle user data and privacy for assistant interactions.
Why this would matter
- Performance: Gemini’s advanced language and generative capabilities could significantly boost Siri’s conversational and reasoning skills.
- Privacy implications: On‑device inference differs from cloud calls to third‑party infrastructure—users and regulators will want clarity on data flows, retention, and anonymization.
- Strategic shift: Partnering with Google would mark a notable change in Apple’s approach of building most core services in‑house.
Questions to watch
- Will Apple provide an opt‑out or local‑only mode for privacy‑sensitive users?
- How will Apple communicate the role of a Google‑based model in Siri?
- What contractual terms and technical guarantees (security, updates, audits) will govern the partnership?
For now this is reporting and should be treated as unconfirmed until Apple issues an official statement. Watch for updates from Apple and major tech reporters for more concrete details.
Further reading: follow Mark Gurman’s reporting at Bloomberg and Apple’s newsroom for official announcements (links open in a new tab).
Discussion: Would you be comfortable if Apple used a Google‑built AI to power Siri — pragmatic progress or a privacy concern?