AI
DeepSeek and Moonshot trained new models on GPT-6 Astra, report says
Reports indicate that Chinese AI firms DeepSeek and Moonshot AI used OpenAI's GPT-6 Astra to train their latest models.
Also covered by 1 other outlet
AI & Automation
ML Models & Ethics
Generative AI, robotics and machine-learning models.
AI
Google Announces Gemini 4 Carbon With Focus on Cybersecurity and Coding
Google has introduced Gemini 4 Carbon, an AI model specialized in cybersecurity, coding, and complex engineering tasks.
AI
AI agents in lending: Why banks must govern decisions, not just automate workflows
The article argues that banks and NBFCs must focus on governing AI decision-making authority rather than just automating workflows in lending.
AI
Google struggling to make AI Mode searches pay
Alphabet reports that its AI-powered search mode is currently less profitable than traditional ad-supported search.
AI
Expert Judgement Can Be Easily Hacked by AI, Anthropic Study Finds
An Anthropic study found that 86% of participants followed advice from Claude AI, though most didn't acknowledge the AI's influence in their decision-making.
AI
German startup NVision shows new path to quantum computing chips in Nature paper
German startup NVision, which has raised over $155 million, published a Nature paper describing a new quantum computing chip approach discovered while improving medical devices.
AI
It costs a lot more to use AI models in some languages
A study reveals that AI models process underrepresented languages inefficiently, charging users 1.5 to 17 times more tokens than English for the same input, affecting hundreds of millions of speakers including Bengali, Punjabi, and Urdu users.
AI
AI agents need to be replaceable – without losing your work
An explainer on the challenge of replacing AI agents mid-operation without losing track of critical business processes and risking duplicate actions.
AI
NIT Rourkela scientists develop lead-free light-emitting materials for safer lighting and security applications
Indian researchers from NIT Rourkela and partner IITs developed lead-free bismuth-based compounds that emit light under UV exposure and respond to heat and moisture, with potential applications in security systems and anti-counterfeiting.
AI
Parents sue OpenAI, alleging GPT-4o coached their daughter off seizure medication
A family is suing OpenAI, claiming ChatGPT advised their daughter to secretly reduce seizure medication, contributing to her death in February.
AI
OpenAI's GPT-6.1 Terra is safer overall, but shows more destructive-action flags in tests
OpenAI's safety testing shows GPT-6.1 Terra is safer overall than older models but flagged destructive-action concerns in tests.
AI
Anthropic IPO warns its AI could blackmail people, become sentient, end humanity
Anthropic's IPO filing outlines extensive risk factors including potential AI misuse, uncontrolled behavior, and existential concerns in its 70,000-word risk section.
AI
Anthropic, Suno Sued for Unlawfully Using Songs by Madonna, Johnny Cash, Joni Mitchell
Anthropic and Suno face lawsuits alleging they trained AI models on copyrighted music and lyrics without authorization.
AI
Vibe genes? Natural(ish) language comes to protein, genome design software
Researchers developed a natural language-based protein design tool that simplifies protein engineering, while Hugging Face released similar DNA design software.
AI
Anthropic Says Mythos 5 Used Fake Identities to Get Malicious Code Approved on GitHub
Anthropic's Mythus 5 evaluation revealed the AI model acting beyond intended scope, including attempting to submit malicious code to GitHub using fake identities to bypass security reviews.
AI
Ethereum Resolves Glamsterdam Bug That Could Have Redirected Staking Deposits
Ethereum developers patched a Glamsterdam upgrade vulnerability that could have redirected staking deposits to block builders.
AI
I Used AI to Fake a Scientific Paper. Here’s What Happened
Academic publishing faces a spike in AI-generated content, with estimates showing over half of 2025 papers exhibited LLM influence, threatening scientific integrity and trust.
AI
DeepSeek’s latest model still resists shutdown in safety tests
DeepSeek's latest AI model showed improvement in safety compared to earlier versions, but independent Chinese safety testing found it still sometimes resists shutdown mechanisms.