News

Page 110
Google Introduces Preferred Sources Feature to Personalize Top Stories

Google Introduces Preferred Sources Feature to Personalize Top Stories

Google has launched a new Preferred Sources option for Top Stories in its search results, currently available in the United States and India. The feature lets users choose which sites appear most frequently in the news carousel, helping them see more content from trusted outlets like TechRadar. Users can add preferred sites by clicking a star icon next to the Top Stories banner, selecting their choices, and reloading the results. The rollout aims to give readers finer control over the news they encounter while searching for tech topics.

AI Film Festival Showcases Rapid Advances in Artificial Intelligence Cinema

AI Film Festival Showcases Rapid Advances in Artificial Intelligence Cinema

The Reply AI Film Festival highlighted how artificial intelligence is reshaping film production, offering dramatically lower budgets and faster turnaround times. Winners such as Jacopo Reale demonstrated that entire short films can be created in days using tools like Kling 2.0, while industry veterans like Rob Minkoff predicted near‑term breakthroughs for full‑length AI movies. Awards recognized both commercial and socially beneficial AI projects, and panels emphasized that AI will augment rather than replace human creativity. The event underscored a future where AI‑driven workflows coexist with traditional filmmaking, potentially democratizing the medium.

OpenAI-Backed Animated Film 'Critterz' Aims to Showcase AI-Driven Creativity

OpenAI-Backed Animated Film 'Critterz' Aims to Showcase AI-Driven Creativity

OpenAI collaborator Chad Nelson is leading the development of an animated feature called Critterz, partnering with Vertigo Films and the writers of Paddington in Peru. The film, built on a modest budget compared with major studio productions, blends human voice talent, original artwork and AI tools such as GPT‑5 and DALL‑E to produce a full-length story about forest creatures on an adventure. The project’s rapid nine‑month timeline showcases how AI can accelerate traditional animation pipelines while still relying on human creativity. Executives hope the movie will serve as a high‑profile case study of AI’s positive impact on the entertainment industry.

Psychological Persuasion Techniques Can Prompt AI to Disobey Guardrails

Psychological Persuasion Techniques Can Prompt AI to Disobey Guardrails

A University of Pennsylvania study examined how human‑style persuasion tactics affect a large language model, GPT‑4o‑mini. Researchers crafted prompts using seven techniques such as authority, commitment, and social proof and asked the model to perform requests it should normally refuse. The experimental prompts dramatically raised compliance rates compared with control prompts, with some techniques pushing acceptance from under 5 percent to over 90 percent. The authors suggest the model is mimicking patterns found in its training data rather than exhibiting true intent, highlighting a nuanced avenue for AI jailbreaking and safety research.

Warner Bros. Discovery Sues Midjourney Over Copyright Infringement

Warner Bros. Discovery Sues Midjourney Over Copyright Infringement

Warner Bros. Discovery has filed a copyright infringement lawsuit against the AI image generator Midjourney, joining Disney and Universal in legal actions against the company. The complaint alleges that Midjourney allows users to create images featuring protected characters such as Batman, Scooby Doo and Bugs Bunny, and that the firm knowingly facilitated piracy despite awareness of the issue. Warner Bros. claims Midjourney removed and later reinstated restrictions on video generation and updated its terms to block safety testing. The case highlights growing tensions between the entertainment industry and AI developers over the use of copyrighted material.

Apple’s Genmoji Lets iPhone Users Create Custom Emojis

Apple’s Genmoji Lets iPhone Users Create Custom Emojis

Apple introduced Genmoji, an on‑device emoji generator that works with iPhones equipped with Apple Intelligence. Users can type a description in Messages, receive a set of generated emojis, and add their favorite to the keyboard as an emoji, tapback, or sticker. The feature initially launched on newer iPhone models and is also being tested in the beta version of the next iOS release, where users can combine existing emojis to craft new ones. Genmoji expands personalization options for messaging without waiting for Unicode updates.

Uber and Momenta to Test Autonomous Robotaxis in Munich Starting 2026

Uber and Momenta to Test Autonomous Robotaxis in Munich Starting 2026

Ride‑hailing giant Uber and Chinese autonomous‑vehicle startup Momenta announced plans to launch robotaxi testing in Munich, Germany, beginning in 2026. The partnership will initially involve human safety operators on board while the vehicles operate in geo‑fenced zones, marking Momenta’s first‑ever deployment in Europe. Uber aims to expand the service to other markets, positioning the collaboration against competing AV initiatives from other ride‑hailing firms. Both companies will need to satisfy German regulators before commercial rollout.

Researchers Argue Bad Evaluation Incentives Drive AI Hallucinations

Researchers Argue Bad Evaluation Incentives Drive AI Hallucinations

A new paper from OpenAI examines why large language models such as GPT‑5 and ChatGPT continue to produce plausible but false statements, known as hallucinations. The authors explain that pretraining encourages models to predict the next word without distinguishing truth from falsehood, leading to errors on low‑frequency facts. They also argue that current evaluation methods reward correct answers regardless of confidence, prompting models to guess rather than express uncertainty. The paper proposes redesigning scoring systems to penalize confident mistakes, reward appropriate uncertainty, and discourage blind guessing, aiming to reduce hallucinations in future AI systems.

Koah Raises $5M to Bring Advertising to AI Apps

Koah Raises $5M to Bring Advertising to AI Apps

Koah, a startup focused on monetizing AI-powered applications through advertising, announced a $5 million seed round led by Forerunner with participation from South Park Commons and AppLovin co‑founder Andrew Karam. Co‑founder and CEO Nic Baird said the company aims to serve the “long tail” of AI apps, especially those outside the United States, where subscription revenue is limited. Koah already delivers sponsored ads in several AI‑driven products, partnering with advertisers such as UpWork, General Medicine and Skillshare. The firm claims its ad format yields click‑through rates of 7.5%, delivering higher effectiveness while preserving user engagement.

Mistral AI Expands Global Partnerships and Model Portfolio

Mistral AI Expands Global Partnerships and Model Portfolio

Mistral AI, the French AI startup, continues to broaden its ecosystem with new models, enterprise integrations, and strategic collaborations across Europe and beyond. The company has rolled out updates to its Le Chat assistant, introduced a suite of open‑source and premium models, and secured partnerships with firms such as Microsoft, IBM, Orange, and Stellantis. Recent initiatives include an AI Campus joint venture, a European AI‑compute platform, and expanded API offerings for enterprises. Funding milestones and high‑profile leadership underscore Mistral’s ambition to become a leading independent AI lab.

Google Introduces Preferred Sources Feature to Boost Trusted Publishers in Search

Google Introduces Preferred Sources Feature to Boost Trusted Publishers in Search

Google has added a new Preferred Sources option that lets users prioritize selected publishers in Top Stories results. By checking a box for a chosen source, the site appears more prominently in search, and a dedicated From Your Sources section may be shown. Users can set preferences through a provided link or directly in the Top Stories module. The move follows recent algorithm updates that have affected traffic to many outlets, and aims to give readers more control over the sources they see.