AI News

News

Page 196
Pinterest Adds Controls to Trim AI-Generated Posts in Home Feed

Pinterest Adds Controls to Trim AI-Generated Posts in Home Feed

Pinterest has introduced a new setting that lets users limit the amount of AI‑generated content appearing in their home feed. Available now on Android and the web, with an iOS rollout planned, the feature lives under a "GenAI Interests" tab in the home feed tuner. Users can toggle the switch off for specific categories that are prone to AI creation, reducing—but not eliminating—such posts. The move reflects Pinterest’s aim to balance human creativity with AI innovation, according to Chief Technology Officer Matt Madrigal.

Ed Zitron Warns AI Bubble Could Burst, Citing Nvidia and Sam Altman

Ed Zitron Warns AI Bubble Could Burst, Citing Nvidia and Sam Altman

In a recent Ars Live discussion, Ed Zitron warned that the AI boom is inflating a market bubble, largely driven by Nvidia’s growth. He argued that a slowdown in that growth could trigger a burst, potentially leading to a broader market depression as investors reassess the sustainability of tech’s hyper‑growth. Zitron also criticized OpenAI CEO Sam Altman, labeling him a con artist who has misled the market. The conversation highlighted concerns over AI hype, the risk of inflated valuations, and the possible fallout for Silicon Valley if the bubble collapses.

Google's Ask Photos Feature Unavailable in Texas and Illinois Amid Biometric Privacy Concerns

Google's Ask Photos Feature Unavailable in Texas and Illinois Amid Biometric Privacy Concerns

Google has confirmed that its AI-powered Ask Photos feature is currently unavailable to users in Texas and Illinois. The company cited ongoing efforts to determine how to expand access, while industry observers link the restriction to recent state settlements over biometric data collection in Google Photos. Both Ask Photos and the related Conversational Editing tool rely on facial recognition, which raises legal challenges under state privacy laws that require explicit consent from subjects captured in photos.

Tesla Reintroduces 'Mad Max' Full Self-Driving Mode Amid Safety Concerns

Tesla Reintroduces 'Mad Max' Full Self-Driving Mode Amid Safety Concerns

Tesla has revived a controversial driving mode called Mad Max for its Full Self-Driving system. Marketed as a faster, more aggressive option than the existing Hurry mode, Mad Max was originally part of the 2018 Autopilot suite. Within hours of its return, the feature was observed ignoring stop signs and exceeding speed limits by more than 15 mph (24 km/h). The U.S. National Highway Traffic Safety Administration has opened an investigation after receiving over 50 reports of traffic safety violations and numerous crashes linked to the system. Critics describe the move as reckless and arrogant, highlighting Tesla’s ongoing safety challenges.

Pinterest Introduces ‘AI Tuner’ to Let Users Reduce AI-Generated Content in Their Feed

Pinterest Introduces ‘AI Tuner’ to Let Users Reduce AI-Generated Content in Their Feed

Pinterest has launched a new feature called the “AI tuner” that lets users dial down the amount of AI‑generated content they see. The tool works on eligible image Pins in categories prone to AI content such as beauty, art, fashion and home decor. Accessible now on Android and desktop, the tuner will roll out to iPhone users in the coming weeks. It sits under Settings → Refine Your Recommendations → GenAI Interests. The move follows Pinterest’s earlier effort to label AI‑modified Pins with a notice in the lower‑left corner and to improve its detection systems.

GZDoom Community Divides Over AI-Generated Code Inclusion

GZDoom Community Divides Over AI-Generated Code Inclusion

A heated debate erupted within the open‑source GZDoom community after project lead Graf Zahl added AI‑generated code to the repository. Critics accused him of violating GPL principles and using unverified snippets, while Zahl defended the move as a harmless shortcut for boiler‑plate tasks. The controversy led to calls for forks, strong backlash on GitHub, and a broader discussion about the role of generative AI in open‑source development.

OpenAI Expands Sora with Storyboarding and Longer Video Limits

OpenAI Expands Sora with Storyboarding and Longer Video Limits

OpenAI announced new features for its AI‑generated video app Sora, adding a web‑based storyboarding tool for Pro users and extending video length limits. Free users can now create videos up to 15 seconds, while Pro users on the web can generate clips up to 25 seconds. The updates aim to attract more professional creators and align Sora with industry practices such as storyboarding, while maintaining its integration with ChatGPT accounts.

Anthropic Introduces ‘Skills’ Feature for Claude AI, Boosting Productivity

Anthropic Introduces ‘Skills’ Feature for Claude AI, Boosting Productivity

Anthropic has launched a new capability called Skills for Claude AI users across its Pro, Max, Team and Enterprise plans. Skills are pre‑configured folders that embed instructions, brand guidelines, legal clauses and other reference material directly into Claude projects, allowing the model to produce consistent, on‑brand output without repeated prompting. The feature is integrated throughout Claude’s ecosystem, including its apps, Claude Code, the Developer Platform and the Agent SDK. At the same time, Anthropic announced the Haiku 4.5 model, a faster and more efficient version of Claude that retains the coding strengths of the earlier Sonnet release.

OpenAI Subpoenas Critics Amid Musk Lawsuit Over Restructuring

OpenAI Subpoenas Critics Amid Musk Lawsuit Over Restructuring

OpenAI has issued subpoenas to several nonprofit groups that have criticized its recent restructuring, including the Midas Project, the San Francisco Foundation, and Encode. The move follows a lawsuit filed by Elon Musk accusing OpenAI of abandoning its nonprofit mission in pursuit of profit. Representatives from the targeted organizations deny any funding or coordination with Musk and claim the subpoenas aim to silence dissent. OpenAI officials assert that the groups are part of campaigns opposing the company's restructuring.

Google-Yale AI Model Uncovers Hidden Cancer Cells, Offering New Treatment Path

Google-Yale AI Model Uncovers Hidden Cancer Cells, Offering New Treatment Path

A joint effort by Google and Yale University has produced a 27‑billion‑parameter AI model, called Gemma, that can identify drugs capable of turning "cold" tumors visible to the immune system. The model evaluated 4,000 compounds, pinpointing both known agents and unexpected candidates. In laboratory tests, a combination of interferon and the AI‑selected drug silmitasertib boosted antigen presentation, effectively exposing previously hidden cancer cells. Researchers say the discovery could lead to earlier detection and new therapeutic strategies for cancers that often evade diagnosis.

Anthropic Launches "Skills" to Boost Claude’s Workplace AI Capabilities

Anthropic Launches "Skills" to Boost Claude’s Workplace AI Capabilities

Anthropic announced a new feature called Skills for Claude, a collection of pre‑built instructions, scripts and resources that let the Claude AI agent perform specific work tasks more efficiently. Companies such as Box, Rakuten and Canva have already used the tool, and users can also create custom Skills for their own workflows across Claude.ai, Claude Code, the API and the Claude Agent SDK. The rollout aims to reduce prompt‑engineering effort and make AI agents more reliable in real‑world business contexts, positioning Anthropic alongside rivals like OpenAI, which recently unveiled its own AgentKit suite.

ByteDance’s Doubao Surpasses DeepSeek to Become China’s Leading AI Chatbot

ByteDance’s Doubao Surpasses DeepSeek to Become China’s Leading AI Chatbot

ByteDance’s Doubao app has overtaken DeepSeek to become the most popular AI chatbot in China, drawing tens of millions of monthly active users. Designed with a friendly avatar and deep integration with Douyin, Doubao offers text, audio, video, image generation and customizable agents. Analysts credit ByteDance’s expertise in viral app design and seamless cross‑platform sharing for Doubao’s rapid growth, while DeepSeek remains a more minimalist text‑only service. The company is now exploring extensions of Doubao into smart glasses, cars and other devices, signalling an ambition to broaden the AI assistant beyond smartphones.