Generative AI

Tilly Norwood's live interview took an unexpected turn when the AI actress suddenly spoke Mandarin, leaving Piers Morgan baffled and the internet divided.

AI is resurrecting the dead in viral videos and personal chatbots, but the trend raises thorny questions about consent, exploitation, and the nature of grief.

Inside a lab where ancient DNA and AI language models hunt for the next antibiotic.

Instagram is suddenly a 1986 time capsule, and ChatGPT is the time machine behind it.

Suno just changed how AI music gets made. The v6 launch is not about better audio. It is about who gets paid.

Social media's latest AI obsession turns selfies into grainy 1980s portraits, and even politicians and cricketers are joining the nostalgia wave.

A quiet API beta hints at DeepSeek's next architectural leap before anyone gets a formal spec sheet.

AI-generated food images on restaurant menus are triggering consumer backlash, vandalism, and new research on how synthetic visuals alter appetite and trust.

Google's WeatherNext 3 uses live satellite data to deliver hourly, high-resolution forecasts with major accuracy gains for rain and snow.

AI-generated ads may look identical to human-made ones, but new research shows they underperform on key effectiveness metrics.

Meta's Muse Spark 1.3 launch marks a claimed leap to parity with OpenAI and Anthropic, backed by independent benchmark scores.

Runway introduces Solaris, a new AI model that generates interactive app interfaces in real time, potentially reshaping how users interact with software.

Zhu Shuhan, 13, used a large language model to build Zai Chang, an AI app that acts as a virtual parent during homework, and earned 18,000 yuan in three days.

Rita Kothari and Arunava Sinha explain why AI cannot replace literary translators, citing data gaps, mechanical processes, and the need for human contextualisation.

A proof by OpenAI's Astra model has mathematicians asking whether their field's value lies in theorems or human understanding.

The co-founders of Accelerated Understanding Inc reveal a physics-first AI model that processes 5 trillion data points, after walking away from a Bezos-backed offer.

ChatGPT's new sticker feature turns personal photos into shareable packs for WhatsApp and iMessage, no design skills needed.

Alibaba has officially launched Wan3.0, an upgraded AI video generation model, just after raising $10 billion to support AI spending.

Alibaba has launched HappyShrimp 1.0, an AI music model that transforms emotional prompts into complete songs, though technical and pricing details are still unknown.

Higgsfield's valuation quadrupled to $5.4 billion in six months after a $400 million Series B, highlighting fierce investor demand for AI-generated marketing and media content.

At MIST 2026, storytellers argued that AI's dominance could create a counter-trend, making live oral storytelling more valuable than ever.

Mirage launched the first AI news network with synthetic anchors streaming live on X, raising urgent questions about trust and the future of journalism.

An AI-generated song from Samay Raina’s show has racked up half a billion views, forcing a reckoning over authorship, taste, and whether the human element in music ever mattered.

Mark Zuckerberg’s new manifesto argues the real AI danger is centralized power, not rogue agents, and proposes personal superintelligence through open-weight models.

A short AI film depicting a zombie apocalypse during Singapore's National Day Parade has sparked intense debate, with creator Joel Boh revealing high audience retention despite criticism.

Meta releases Muse Glimmer, a compact open-weight model that brings agentic AI to consumer hardware, reigniting the open-source versus closed AI debate.

China has released OneGenome, a free AI tool that applies clinical reasoning to genomic data, offering hope for faster diagnosis of rare genetic diseases.

Researchers at Stanford and the Arc Institute have used generative AI to design fully functional bacteriophages from scratch, a first in synthetic biology that could revolutionize phage therapy but also raises biosecurity alarms.

The Bayreuth Festival’s 150th anniversary staging of Wagner’s Ring cycle used AI to generate real-time visuals, but the result was a disjointed, unmoving spectacle that left singers stranded behind a screen and drama in tatters.

DeepSeek's V4 Flash model disrupts the AI market with 99% lower costs than Anthropic's Claude Opus 4.8, intensifying the industry price war and fueling the commoditization of AI.