AI News

shopify canvas build online store chatting with ai a blueprint with a model shop standing on it

Shopify Debuts Canvas, a Way to Build Online Stores by Chatting With AI

Shopify has launched Canvas, a desktop workspace where merchants build or redesign their online store by chatting with Sidekick, its AI agent, while seeing every page at once. We explain how Shopify Canvas works step by step, what Sidekick does behind the scenes, how it builds on Shopify’s earlier AI tools, the limits at launch, what it means for agencies, and whether your business should try it now.

Read more
ai writing tells opus 5 5 this matters graphite a whack a mole cabinet with moles popping up

Opus 5.5 Loves to Tell You ‘This Matters’ (and Other AI Writing Tells)

A study from the marketing agency Graphite finds that Claude Opus 5.5 uses the phrase “this matters” 116 times as often as human writers, and still has 2,548 AI writing tells despite nearly abandoning the em dash. We explain how the study worked, Opus 5.5’s biggest tells, how they differ from Opus 5 and GPT-6 Astra, why tells keep returning, the study’s limits, and how to edit them out of a draft.

Read more
openai firings wsj names three safety researchers metr a meerkat standing sentry on its burrow mound

OpenAI Cuts Ties With 3 Safety Researchers: WSJ Names Them and a METR Link Emerges

The Wall Street Journal has named the three researchers OpenAI dismissed for allegedly sharing confidential information with an outside AI safety organisation: Jasmine Wang, Tomek Korbak and Mikita Balesni. One of them was OpenAI’s technical contact for METR’s independent investigation of the Hugging Face hack. We set out who the three are, what the METR link does and does not show, how outside investigators get access to a lab, and what the case means for oversight and for businesses using OpenAI.

Read more
grok venezuela trump asked chatbot maduro capture a firework bursting into stars above a launch tube

Musk’s AI Chatbot Grok Reportedly Encouraged Trump to Capture Venezuela’s President

TIME reports that Donald Trump spent hours asking Elon Musk’s Grok chatbot questions in December 2025, including how Venezuelans would react if the US captured Nicolas Maduro. We set out exactly what TIME reported and how the story changed in the retelling, what Venezuelans actually thought, how Grok is used by the Pentagon, and what the episode teaches about AI in decision-making.

Read more
primary bot grok 4 7 base model proactive grok bot a robot butler holding out a serving tray

xAI Rolls Out Grok 4.7 as Base Model and Introduces Proactive Primary Bot

SpaceXAI has given Grok Bot a proactive Primary Bot that spots work and offers to handle it before you ask, and Grok 4.7 is now the base model across the Fast, Expert, Build and Heavy modes in the Grok apps. We cover how Primary Bot works, what free suggestions really cover, Grok 4.7’s benchmark gains, what is still undocumented, and what it means for businesses.

Read more
suno speech ai voice music spoken word beta a lyre with taut strings and rising music notes

AI Music Maker Suno Now Generates Spoken Words

Suno has launched Speech in public beta, a model that generates spoken-word audio and original background music together in one track. We explain how Simple and Advanced modes work, what Suno has and has not said about credits, voices and rights, why the AI music company is diversifying, and what it means for businesses using synthetic voice.

Read more
ai voice production onepin agent 10 languages a mixing desk with rows of faders

Onepin AI Voice Production Agent Supports 10 Languages for Large-Scale Projects

Onepin, from speech-evaluation company Podonos, is an AI voice production agent that routes each line of a script to a text-to-speech engine, scores it for naturalness, clarity and pronunciation, and fixes what fails, in 10 languages. We explain the pipeline, the pricing per validated hour, where Onepin’s own claims disagree, who it suits and what UK teams should check.

Read more
openai safety researchers fired leak wsj a ejector seat blasting upward on a rocket plume

OpenAI Cuts Ties With 3 Safety Researchers, WSJ Reports

OpenAI has parted ways with three members of its safety team who allegedly shared confidential information with an outside AI safety organisation, The Wall Street Journal reported on 1 October 2026. We set out what is known and not known, the 2024 precedents, the run of agent incidents behind the story, what whistleblower law protects in California and the UK, and what it means for businesses using OpenAI.

Read more
mai playground microsoft mai voice transcribe models a geodesic jungle gym climbing frame

Microsoft Adds MAI-Transcribe and MAI-Voice Models to MAI Playground

Microsoft has added three audio models to MAI Playground: MAI-Transcribe-2-Streaming for real-time transcription and the MAI-Voice-2.1 and MAI-Voice-2.1-Flash text-to-speech models, plus a voice assistant demo called Chatter. We cover what each model does and costs, how the streaming model ranks on Artificial Analysis, how fast Microsoft’s voice line has moved and what UK developers should check.

Read more
griffin lite tavus real time video ai preview a retro videophone with a face on its screen

Tavus Launches Limited Preview of Griffin Lite and Updates Developer Platform

Tavus has opened a limited research preview of Griffin Lite, a real-time video AI that sees, hears and talks back with a generated face and voice. In Tavus’s study, 26 of 54 people thought it was human after a one-minute call. We explain how it works, what the NVIDIA VideoFDB scores show, what changed on the Tavus developer platform in September, and the deception risks Tavus itself flags.

Read more
CHAT