Notizie — Pagina147

Anthropic is rolling out an integration that links its Claude AI assistant with Microsoft 365 tools, including Teams, Outlook, SharePoint, and OneDrive. The connector lets Claude pull content from Word documents, chat threads, and email conversations, delivering contextual answers without manual uploads. Available now for Claude Team and Enterprise users, the feature requires admin activation. Anthropic also introduced enterprise search in Claude, enabling broader data retrieval across corporate sources. The integration relies on Anthropic’s open‑source Model Context Protocol (MCP), which Microsoft plans to adopt widely across its platforms.Leggi di più

A recent study by Graphite, using Common Crawl data and AI‑detection tools, determined that more than half of newly published English‑language web articles are now written by artificial intelligence. While the volume of AI‑generated content has plateaued, most of it fails to rank well in Google search or appear in ChatGPT responses, indicating that human‑written pieces still dominate visibility. The findings highlight a shift in how publishers, marketers, and content farms produce material, as well as ongoing concerns about quality, SEO performance, and the future role of AI in online publishing.Leggi di più

Spotify announced a collaboration with the recording divisions of Sony, Universal and Warner to create new artificial‑intelligence‑driven music offerings. The partnership emphasizes artist‑centric agreements that protect copyright, ensure fair compensation, give creators a choice to participate, and strengthen fan‑artist engagement. Spotify also highlighted its ongoing AI tools, including an AI‑DJ feature and a spam filter aimed at reducing unauthorized AI‑generated tracks on the platform.Leggi di più

Meta AI now appears in Facebook, Instagram, Messenger, and WhatsApp, offering suggestions and comment summaries that many users find intrusive. While the feature cannot be removed entirely, each app provides a way to mute the AI chat and turn off its automatic functions. By accessing the chat settings or privacy options within each platform, users can silence Meta AI, hide its messages, and stop comment summaries from appearing on their posts. These steps give users control over the AI presence without needing to uninstall any apps.Leggi di più

Moderators on Reddit have raised alarms after the platform's AI feature, Reddit Answers, suggested high‑dose kratom and even heroin as treatments for chronic pain. The unsafe recommendations prompted concerns about the tool's ability to handle medical queries. Reddit responded by updating the AI to hide related answers on sensitive topics, but moderators say the change does not address broader risks of AI‑generated health advice and leaves them without effective controls.Leggi di più

Microsoft is testing a suite of new AI-driven capabilities for Windows 11 through its Insider program. The updates include Copilot taking over the taskbar search area, answering plain‑language questions about system settings, and linking to the appropriate Settings pages. Additional features such as Copilot Connectors, document‑creation shortcuts, File Explorer AI actions, and Vision‑enabled document scanning are being trialed. While none of the changes are yet part of the publicly released Windows 11 25H2 update, the company appears to be applying lessons learned from earlier rollout challenges.Leggi di più

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.Leggi di più

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.Leggi di più

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.Leggi di più

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.Leggi di più

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.Leggi di più

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.Leggi di più

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.Leggi di più

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.Leggi di più

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.Leggi di più

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.Leggi di più

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.Leggi di più

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.Leggi di più

Spotify announced a partnership with the three major record labels—Sony, Universal and Warner—along with digital‑rights groups Merlin and Believe to develop AI‑driven music tools that prioritize artists. The initiative, outlined in a blog post, emphasizes responsible AI use, fair compensation, and direct artist‑fan connections. It follows Spotify’s recent effort to clean up unauthorized AI‑generated tracks on its platform and builds on its own AI features such as the AI DJ, personalized daylist, and AI‑generated playlists. The company frames the collaboration as a proactive step to shape the future of music innovation responsibly.Leggi di più

OpenAI announced two major updates to its AI video generation tool Sora 2. All users can now create videos up to 15 seconds, while Pro subscribers can generate clips up to 25 seconds and gain access to a new Storyboard tool that lets them combine multiple scenes. The upgrades aim to give users more creative flexibility and strengthen Sora 2’s position as a leading AI video platform.Leggi di più