AI News
Daily Tech Reader - AI Edition
Apple
Apple
Apple Sets Sept. 9 Event for iPhone 18 Pro and Foldable iPhone Ultra
Apple is expected to unveil the iPhone 18 Pro lineup, the foldable iPhone Ultra, and Apple Watch Series 12 at its Sept. 9 event, marking CEO John Ternus’s first keynote since taking over.
Sources: MacRumors · Digital Trends · ASO World
iOS 27 Developer Betas Bring Gemini‑Powered Siri Overhaul
Apple released developer betas of iOS 27 featuring a rebuilt Siri powered by Google Gemini models, enabling contextual conversations, on‑screen awareness, and multi‑step actions across apps.
Sources: PhoneArena · Apple Newsroom · MacRumors
iPhone 18 Pro to Feature A20 Pro Chip and 12GB RAM for On‑Device AI
Apple’s A20 Pro chip built on a 2nm process will power the iPhone 18 Pro, with 12GB of RAM supporting expanded on‑device Apple Intelligence features and a larger Neural Engine.
Sources: MacRumors · Yahoo Tech
Microsoft
Microsoft Adds Invisible Watermarks to Copilot AI Content Under EU AI Act
Microsoft has begun applying invisible watermarks to AI‑generated audio and video created in Microsoft 365 Copilot, meeting compliance requirements under the EU AI Act’s Code of Practice.
Sources: Microsoft Learn · Dynamics Edge
Claude Opus 5 Now Available in Microsoft 365 Copilot
Anthropic’s Claude Opus 5 became available inside Microsoft 365 Copilot in late July, giving enterprise users access to a second frontier‑level model alongside GPT‑4.1.
Sources: Microsoft Community Hub · Dynamics Edge
Copilot Domain Exclusion Reaches General Availability
Microsoft released Domain Exclusion for Microsoft 365 Copilot, allowing administrators to block specific websites from being used as grounding sources in Copilot responses.
Sources: Microsoft Community Hub · Microsoft Learn
Claude
Partial Claude Outage Affects Mythos 5.1, Fable 5.1, and Opus 5
Anthropic reported a partial outage affecting several Claude models beginning around 9:41 AM ET, disrupting API calls and the Claude web interface. The company said it has identified the cause and is working on a fix.
Sources: BleepingComputer · CyberSecurityNews · DesignTAXI Community
Anthropic Adds Watermarking to Claude Outputs Under EU AI Act
Anthropic began watermarking Claude outputs in August to comply with the EU AI Act’s Code of Practice, with a detection API available for organizations needing to verify Claude‑assisted text.
Sources: Anthropic · Releasebot
Claude Code Promo Ends Sept. 14; Permanent Limit Increase Follows
Anthropic said its summer Claude Code usage promotion will end Sept. 14, replaced by a permanent 25% increase over baseline limits.
Sources: ExplainX.ai · Releasebot
ChatGPT
OpenAI Launches GPT‑6 Astra, Calling It Its Most Capable Model to Date
OpenAI released GPT‑6 Astra, describing it as its most powerful model and the first to reach its Critical cybersecurity threshold. The rollout begins with Daybreak enterprise customers, with broader access expected soon.
Sources: TechCrunch · VentureBeat · Bloomberg · CNBC
OpenAI Restricts Astra’s Advanced Cyber Capabilities to Vetted Users
OpenAI said Astra can autonomously identify and exploit zero‑day vulnerabilities and scored 100% on ExploitBench. The company is limiting advanced cyber features to its Daybreak program while releasing the general model publicly.
Sources: CNBC · TechCrunch · Axios
U.S. Justice Department Supports OpenAI in New York Times Copyright Case
The Justice Department filed a statement of interest backing OpenAI’s fair‑use defense in the lawsuit brought by The New York Times, ahead of upcoming summary‑judgment motions.
Sources: AI Weekly · TechCrunch
Gemini
Google Releases Gemini 3.8 Flash With Improved Reasoning and Coding
Google launched Gemini 3.8 Flash, calling it its most capable Flash model for coding and multi‑step agentic tasks. Pricing remains at $0.75 per million input tokens.
Sources: CNBC · Tech‑Insider
Alphabet Leans on Gemini 3.8 Flash as Part of AI Recovery Strategy
Google’s rapid Flash model iteration comes as Alphabet emerges from a prolonged market downturn, with Gemini’s coding and agentic improvements forming a key part of its AI narrative.
Sources: CNBC · Tech‑Insider
Google Expands Gemini Lineup With Spark, Daily Brief, and Gemini Live
Google introduced new Gemini‑based services including Spark, a persistent background agent; Daily Brief, an AI‑curated morning summary tool; and Gemini Live, which provides real‑time screen and camera analysis.
Sources: Blog.mean.ceo · Google AI
China AI
South Korea Announces $919 Billion Sovereign AI Program
South Korea unveiled a sovereign AI plan targeting 8.4GW of data‑center capacity by 2029, with initial deployments across SK Group, GS Group, and Naver and a government‑run competition to select a national foundation model.
Sources: AI Weekly · SemiAnalysis
Meta Ends AI‑Use Performance Scoring After Legal Pushback
Meta is discontinuing AI‑adoption metrics tied to employee performance reviews following legal challenges, while rolling out Hatch, a new internal AI agent, across the company.
Sources: Wired · AI Weekly · TechCrunch
Big Story
OpenAI’s GPT‑6 Astra Marks Major Step in AI Capability and Cyber Risk
OpenAI launched GPT‑6 Astra, its first model rated at the Critical cybersecurity tier, capable of autonomous zero‑day discovery and exploitation. The company is gating advanced features while expanding enterprise access.
Sources: VentureBeat · TechCrunch · Bloomberg · CNBC