AI in News – Curated Headlines That Matter Today

Explore Stories by Date

29 Jun 2026

Amazon seeks cheaper AI alternatives as Anthropic shifts to token-based pricing next year

Amazon is exploring OpenAI and other options after a renegotiated contract with Anthropic will introduce token-based pricing that could substantially increase costs. The shift threatens Amazon's heavy reliance on Claude AI for internal tools like coding agent Kiro, workplace assistant Quick, and Alexa for Shopping, signaling a deepening rift between the once-inseparable AI partners.

2 Sources

New training method doubles accuracy in spotting AI-generated faces, study shows

Researchers at Australian National University developed a training method that nearly doubled people's ability to detect AI-generated faces, with some participants achieving near-perfect accuracy. Instead of looking for visual glitches, the approach teaches people to recognize broader patterns like symmetry and memorability—qualities where AI faces drift toward statistical averages while real human faces embrace distinctive imperfections.

4 Sources

Apple Rushes Security Updates Early as AI-Powered Hacking Compresses Patch Windows

Apple released iOS 26.5.2 and macOS updates weeks ahead of schedule, patching 29+ vulnerabilities originally planned for version 26.6. The company told Reuters it's adapting to AI-driven hacking tools that shrink the window between vulnerability disclosure and exploitation. While none of the flaws were actively exploited, the shift marks a fundamental change in how Apple deploys security fixes.

11 Sources

AI hiring tools show racial bias against Black and Asian applicants, Stanford study reveals

A Stanford University study analyzing over 4 million job applications found that AI hiring tools displayed racial bias on a job-by-job basis, with 26% of Black applicants and 15% of Asian applicants facing biased algorithmic recommendations. The research also uncovered systemic rejection patterns, where 4% of candidates applying to 10 positions were rejected by AI across all jobs—suggesting an algorithmic monoculture problem.

2 Sources

Meta Brain2Qwerty v2 decodes typed sentences from brain scans at 61% accuracy without implants

Meta unveiled Brain2Qwerty v2, an AI system that converts brain activity into text without surgery, achieving 61% word accuracy using magnetoencephalography scanners. The non-invasive brain-computer interface trained on 22,000 sentences from nine volunteers marks a leap from single-digit accuracy in previous methods, though it requires room-sized equipment and can't operate in real time.

7 Sources

Anthropic Claude launches on Microsoft Azure with NVIDIA GB300 Blackwell Ultra GPUs

Anthropic's Claude AI models are now generally available in Microsoft Foundry on Azure, powered by NVIDIA GB300 Blackwell Ultra GPUs. This marks Anthropic's first NVIDIA deployment, enabling Azure customers to build autonomous and domain-specific AI agents with enhanced inference performance and reduced costs.

4 Sources

Cursor launches iOS app so developers can guide AI coding agents from their phone

Cursor released a mobile app for iOS that lets developers spin up coding agents from phone and manage them remotely. The launch signals a shift toward mobile-first AI-driven development, with even Anthropic's Boris Cherny noting most of his coding now happens on mobile. The move comes as SpaceX's $60 billion acquisition of Cursor progresses.

3 Sources

Google Meet's Gemini AI note-taker expands to Pro and Ultra subscribers at $19.99 monthly

Google Meet now offers its Gemini-powered Take notes for me feature to individual AI Pro and Ultra subscribers. The AI-powered note-taker transcribes calls, generates summaries, and tracks action items automatically. Previously limited to Google Workspace business customers, the tool saves notes to Google Docs and sends email recaps after meetings conclude.

6 Sources

AI Employees Undermine Work Quality as Companies Rush to Deploy Digital Coworkers

New research reveals managers catch 18% fewer errors when AI agents are framed as employees rather than tools. Nearly a third of companies now treat AI agents as coworkers, with 23% listing them on org charts. The findings expose how branding AI as employees creates accountability gaps and undermines the very productivity gains these tools promise to deliver.

2 Sources

DeepSeek open-sources DSpark framework to speed up LLM inference by up to 85%

DeepSeek released DSpark, an open-source framework using speculative decoding to accelerate AI responses by 60-85%. Released under MIT license with DeepSpec training code, the system tackles costly AI deployment challenges. The framework works across multiple open-weight models including Qwen and Gemma, potentially reshaping how developers deploy large language models.

2 Sources

Mirendil raises $200 million to build self-improving AI that accelerates scientific discovery

Former Anthropic researchers Behnam Neyshabur and Harsh Mehta have launched Mirendil with $200 million in seed funding at a $1 billion valuation. The startup aims to democratize access to AI that improves AI—technology that major labs currently keep proprietary. Backed by Andreessen Horowitz, Kleiner Perkins, and Nvidia, Mirendil wants to put recursively self-improving AI systems into the hands of independent researchers to accelerate scientific discovery.

2 Sources

Cloud AI hits capacity limits as Google tells Meta to ration usage, making local AI essential

Google told Meta in March it couldn't supply enough Gemini computing capacity, forcing the company to ration token usage and delay internal projects. The incident reveals cloud AI infrastructure constraints even for tech giants with nine-figure budgets. Meanwhile, local AI solutions are advancing rapidly, with new AI-specific hardware and models like Gemma 4 enabling on-device processing that offers privacy, cost savings, and independence from cloud providers.

4 Sources

TIDAL cuts royalties for AI-generated music while labeling tracks with new AI badge system

Music streaming service TIDAL introduced a policy that demonetizes fully AI-generated music, preventing these tracks from earning royalties or direct-to-fan sales revenue. Starting July 15, the platform will label 100% AI-generated content with an AI badge and remove tracks that impersonate artists, joining competitors like Spotify, Deezer, and Apple Music in addressing the flood of synthetic content.

8 Sources

AI Agents Face Critical Identity Security Gaps as Attackers Exploit Authentication Flaws

AI agents are creating unprecedented identity security challenges that traditional access controls cannot address. Recent attacks demonstrate how agents with legitimate credentials can be hijacked to execute malicious code without triggering alerts. Security leaders report 88% of organizations have experienced AI-related incidents, while agents operate with overprivilege at machine speed across production environments.

4 Sources

AI Regulation Gains Strong Bipartisan Support as 68% of Voters Demand Safety Reviews

A new AI Policy Institute poll reveals overwhelming bipartisan support for tighter AI regulation, with 68% of likely U.S. voters favoring mandatory formal safety reviews for advanced AI systems before public release. The survey shows Republicans are more enthusiastic than Democrats about government-led safety testing, marking a shift in attitudes as concerns mount over AI data centers and infrastructure.

3 Sources

Get all updates into your inbox 💌, Stay Up-to-date for Free 🥳

Don’t drown in AI news. We cut through the noise - filtering, ranking and summarizing the most important AI news, breakthroughs and research daily. Follow topics that matter to you and stay ahead.

Explore

News Categories

Technology Business Policy Startups Health Science Entertainment

Terms Privacy Content Contact Us