The Chinese startup Synyi AI is piloting its first AI-powered medical clinic in Saudi Arabia. At the clinic, located in the eastern part of the country, a virtual doctor called "Dr. Hua" handles diagnoses and drafts treatment recommendations, which are then reviewed and signed by a human physician. The system analyzes symptoms, images, and patient data collected by human assistants. For now, Synyi AI focuses on respiratory illnesses, covering around 30 conditions such as asthma and pharyngitis. The company plans to expand the service to about 50 diseases in the future, including gastroenterological and dermatological conditions. The pilot is designed to generate data needed for regulatory approval in Saudi Arabia. Synyi, which is backed by Tencent, aims to open more clinics and broaden the AI's range to additional diseases. Saudi Arabia is the company's first international market.
YouTube is rolling out a new ad tool called "Peak Points," which uses Google's Gemini AI model to identify the most engaging moments in videos and place ads immediately after them. The goal is to boost ad visibility and click-through rates. According to YouTube, Gemini analyzes elements like video frames and transcripts to pinpoint these high-attention segments. "Peak Points" is currently in a pilot program and is slated for wider release later this year. The feature was announced at the YouTube Brandcast event in New York.
OpenAI is rolling out its GPT-4.1 model to ChatGPT, making it available outside the API for the first time. Plus, Pro, and Team users can now select GPT-4.1 from the "More models" drop-down menu, while Enterprise and Edu users are expected to get access in the coming weeks. GPT-4.1 mini replaces GPT-4o-mini and is now available to all ChatGPT users.
OpenAI says GPT-4.1 is particularly strong when it comes to programming tasks and following instructions precisely. In our tests, the model is noticeably less "chatty" than GPT-4o—not being overly talkative can be useful—and its performance is otherwise on par with the most recent GPT-4o model. Compared to older versions of GPT-4o, GPT-4.1 also feels more thorough.
