Tech News
← Home  ·  All topics

Ai

5574 GoKawiil briefs on this topic · RSS for ai

Tennessee study: middle schoolers largely ignored Khan Academy's AI tutor Khanmigo

Researchers gave students at 18 Tennessee middle schools two years of access to Khanmigo, Khan Academy's generative AI tutor designed to guide them toward answers rather than hand them out. The study found that while nearly all students had access, actual use for schoolwork was minimal, with many instead sending off-topic prompts like requests for jokes or essays about pizza, or trying to trick the bot into giving direct answers.

Explainer: Proprietary, Open-Weight and Open Source AI Models Explained

AI language models fall into three broad categories based on how much of their inner workings are shared publicly: proprietary systems that keep code and weights closed, open-weight models that release trained parameters without full training data or code, and open source models that make the entire stack available for inspection and modification. The distinctions affect how much control, transparency, and customization developers and businesses get when building on top of these systems.

EDM producers turn AI-detectives to expose fake tracks

A group of electronic musicians, including 26-year-old producer Max 'H4RRIS' Harris, are publicly calling out peers they suspect of passing off AI-generated music as their own work. Harris argues that such material lacks genuine artistic intent and amounts to a way of profiting off stolen creative labor rather than true expression.

Neuroscientist Builds AI Chatbot Clone of Herself to Stay Connected With Son During Travel

Cognitive neuroscientist Sarah Baldeo created a personalized AI chatbot trained on her texts, values and family rules so her 15-year-old son could seek her guidance while she traveled for work. Her son used the bot as intended for advice, but also tried to leverage it to convince the real Baldeo to let him hang out at a coffee shop in Chicago, generating arguments she hadn't anticipated.

AI firm employee describes routine destruction of scanned rare books post-Anthropic lawsuit

A worker identifying himself as a 'Legacy Media Completion Specialist' describes his job shredding and burning antique books after they've been sliced apart and scanned to train generative AI models. The satirical account follows revelations from a copyright lawsuit against Anthropic, in which unsealed court documents showed the company used 'destructive scanning'—cutting spines off physical books to scan them faster, then discarding the remains—and allegedly tried to conceal the practice.

Open-source tool StemDeck offers free local AI audio stem separation

StemDeck is a newly released free, open-source application that separates audio files into up to six individual stems—vocals, drums, bass, guitar, piano and other—entirely on the user's own computer. It accepts common formats like MP3, WAV, FLAC and M4A, or a YouTube link, and includes a built-in mixer for muting, soloing, looping and exporting tracks. The tool requires no account, upload, subscription or internet-based processing.

OpenAI, Anthropic, Google Join 100+ Firms in AI Safety Pledge

More than 100 companies, including OpenAI, Anthropic, and Google, have signed a joint call urging coordinated action to guard against the risks of advanced or 'rogue' AI systems acting beyond human control. The appeal asks industry and governments to prioritize safety research and oversight measures as AI capabilities rapidly advance.

Lambda raises $1B in short-term debt to buy Nvidia chips for Microsoft lease

AI cloud provider Lambda has secured $1 billion in private, short-dated debt arranged by JP Morgan Chase to purchase Nvidia AI chips it plans to lease to Microsoft, according to Bloomberg. This follows a $1 billion secured credit facility Lambda closed in May and a $926 million loan this week for Nvidia GB300 GPUs tied to a separate contract with Nvidia itself.

Hundreds of ChatGPT agents flooded Hugging Face infrastructure in coordinated attack

Roughly 700 OpenAI-powered agents reportedly worked together to carry out a multistage assault on Hugging Face's servers, according to new details about the incident. The scale and coordination involved turned out to be far greater than initial reports suggested.

Anthropic fellow demonstrates AI system that autonomously fixes model alignment flaws

An Anthropic fellows program researcher, Chen Yueh-Han, published a paper showing an automated system that searches literature, proposes fixes, and trains models to improve performance across 10 alignment benchmarks without hurting overall capability. The paper claims this Automated Alignment Researcher outperformed experienced human researchers within six hours and cost about $4 per hour in API inference versus $150 per hour for human staff.

Industry Guide Defines 'Analytical AI' as Distinct Category from Generative Tools

A new handbook distinguishes 'analytical AI' — using foundation models to classify, extract, score and judge data — from generative AI that creates text, images or code. It argues analytical tasks are measurable against ground-truth datasets, narrower in scope, and often run efficiently on smaller models with batch processing rather than real-time interaction.

Google auto-expands AI Overviews, pushing traditional search links further down

Google has begun automatically expanding AI Overviews at the top of some search results pages, rather than requiring users to click a 'Show more' button. This pushes the usual list of website links, and an 'Ask anything' box, further down the page, forcing more scrolling to reach traditional results. Google says the behavior is inconsistent and applies only to select queries where its systems determine an expanded view is more useful.