Google Announces AI Models and Services
Google announced the launch and testing status of its Gemini models.
It also introduced tools to identify AI-generated content and detect scams.
The announcements covered medical research, game creation, and an AI response from a satellite.
Gemini Model Launch and Testing
Google said it launched Gemini 4 Argon worldwide for complex coding, financial analysis, and legal work.
The rollout includes Google AI Studio, the Gemini API, Google Cloud, Search, and the Gemini app. Gemini 4 Argon for Robotics is also used to reason about the physical world.
Google announced that the US government and select trusted cyber defenders are testing Gemini 4 Carbon before its release through the Fairwind Program.
Identifying AI-Generated Content
Google said it is expanding SynthID Detector worldwide in English.
The tool can check whether images, videos, and audio are AI-generated.
Scam Detection in Google Messages
Google said it is introducing an optional Enhanced protection mode in Google Messages.
The mode uses Private AI Compute and cloud-based Gemini models to detect sophisticated conversational scams.
AMIE Research and Playground
Google said its AMIE research involved collecting patient histories and generating clinical summaries before urgent care appointments.
According to Google, AMIE helped clinicians prepare in 75% of cases and matched physicians' final differential diagnoses 90% of the time.
Google said Playground, an experimental gaming platform, lets users create, remix, and share single-player and multiplayer games without coding, using simple text prompts.
Project Suncatcher Satellite Response
Google said a Project Suncatcher prototype satellite sent the first AI response generated in space.
The response was generated on a Google TPU aboard the satellite. It was a haiku composed on board by Gemma 4.
Availability and usage conditions vary by feature.