Nachrichten — Seite126

Recent observations of code references labeled "gpt-5-1-thinking" have sparked speculation that OpenAI may soon roll out a new version of its flagship chatbot, dubbed ChatGPT 5.1. The prospective upgrade is expected to expand the model's context window, improve long‑term memory, tighten factual grounding, and offer smoother multimodal interactions that blend text, images, and audio. Users also hope for more reliable automation tools and a steadier conversational style that remembers personal preferences without repeated prompts. While details remain limited, the buzz reflects anticipation that the next iteration will address several lingering shortcomings of GPT‑5.Weiterlesen

The Wikimedia Foundation, which operates Wikipedia, has asked major AI developers to stop scraping the encyclopedia’s data for free and instead pay to use its Enterprise API. The nonprofit argues that AI models need high‑quality, human‑curated information, but the massive traffic and server load caused by unsanctioned scraping threaten Wikipedia’s sustainability. While some companies, including Google, previously reached commercial agreements, many have not responded to the request. The foundation’s appeal adds to a broader movement by online publishers seeking compensation for the use of their content by AI systems.Weiterlesen

Google is expanding its Gemini AI assistant to the Google TV Streamer, launching Gemini for TV today with a rollout over the coming weeks. The new voice‑controlled feature adds robust AI capabilities, offering personalized viewing suggestions and contextual information about programs. The move follows earlier hints from The Android Show and aligns with recent early‑access releases of Gemini for Home on Google Home devices.Weiterlesen

Google has upgraded its NotebookLM mobile app, adding the ability to generate flashcards and quizzes directly from user‑provided source material. The new features sit alongside the existing Audio Overviews tool and come with customizable difficulty and focus options. At the same time, the app’s chat function has been enhanced with a larger context window and faster multiturn capabilities, making interactions more natural and accurate. While the update marks a significant step toward parity with the web version, some advanced features remain unavailable on mobile.Weiterlesen

Google has introduced Nano Banana 2, the next iteration of its AI image generation model, as part of a preview within the Gemini app. The new model emphasizes multi‑step planning, self‑correction, and finer control over angles, viewpoints, and text clarity. Early previews show cleaner lines, sharper angles, and fewer typical AI artifacts. Internally known as GEMPIX 2, Nano Banana 2 also appears in experimental tools like Whisk Labs, signaling Google’s broader push to embed creative AI across its ecosystem.Weiterlesen

Researchers have found that artificial intelligence models keep memorized facts and reasoning abilities in separate parts of their neural networks. By analyzing the loss landscape, they discovered that memorized items create sharp spikes while reasoning produces smoother curves. The study also explored early techniques for removing specific data from models, noting that complete elimination cannot yet be guaranteed. These insights could guide future efforts to manage and protect sensitive information in AI systems.Weiterlesen

Google has begun rolling out its new Gemini voice assistant for Google Home smart speakers and displays, positioning it as the default assistant across its ecosystem. While a recent blog post praised the rollout and highlighted a user’s positive experience, the accompanying FAQ revealed recurring problems, including misinterpreted commands and unreliable device control. Access to Gemini requires enrollment in an Early Access program that launched in late October, and the rollout is expected to be gradual, potentially not reaching all users until next spring. Early feedback suggests the generative‑AI‑driven assistant still struggles with basic smart‑home tasks.Weiterlesen

Google announced that its Gemini generative AI is being rolled out to the Google TV Streamer, where it will replace Google Assistant. The new system lets users ask natural language questions about movies, TV shows, education, DIY projects and more, using the remote microphone button. The rollout begins this week for users aged 18 and older and will expand to additional streaming devices such as select TCL and Hisense models and the Walmart Onn 4K Pro. The move is part of Google’s broader plan to shift its voice ecosystem from Assistant to Gemini across all platforms.Weiterlesen

Major technology firms are committing unprecedented sums to AI‑related data center projects, with Oracle, SoftBank, Meta and others pledging hundreds of billions of dollars. While the scale of funding reflects optimism about AI growth, surveys show many companies remain cautious, and industry leaders warn that physical infrastructure—especially power and space—could become a bottleneck. The juxtaposition of massive financial commitments and lingering market uncertainty highlights the complex dynamics shaping the emerging AI ecosystem.Weiterlesen

Gamma, the AI‑powered platform that creates presentations, websites and social media posts, announced a $68 million Series B round that lifts its valuation to $2.1 billion. Led by Andreessen Horowitz, the round also includes participation from Accel, Uncork Capital and other existing investors. Co‑founder and CEO Grant Lee said the company has reached $100 million in annual recurring revenue on a profitable basis and now serves roughly 70 million users. The funding will support further product development and global expansion.Weiterlesen

The Wikimedia Foundation announced that it expects AI developers to access Wikipedia content through its paid Wikimedia Enterprise platform rather than scraping the site. The move aims to ensure proper attribution, reduce server strain, and support the nonprofit’s mission as human page views decline. The foundation emphasized that generative‑AI tools should credit Wikipedia contributors and direct users back to the source, warning that continued scraping could impact volunteer engagement and donor support.Weiterlesen

A new analysis published in Nature Communications examines the environmental footprint of the rapid AI‑driven data‑center expansion in the United States. While tech giants pour billions into new facilities, the study finds that location dramatically influences carbon and water impacts. States such as Texas, Montana, Nebraska and South Dakota emerge as optimal sites because they balance lower water scarcity with cleaner electricity grids. In contrast, traditional hubs like Virginia and California face growing pressure on energy and water resources, raising doubts about the feasibility of industry net‑zero pledges.Weiterlesen

Google has integrated its Gemini generative‑AI model into Google Photos via a new “Ask” button. The feature lets users query their own photo collections—rather than the web—for specific images, delivering results in seconds. A user with a library of nearly 172,000 photos demonstrated the tool by asking for pictures of a bay window with shades, receiving dozens of relevant shots and even contextual commentary. The rollout showcases Google’s push to embed Gemini across its apps and highlights how AI can streamline everyday tasks like locating a single photo in a massive personal archive.Weiterlesen

Google Maps announced a suite of AI-driven features powered by Gemini models, aimed at both developers and everyday users. The rollout includes a Builder Agent that generates map‑based prototypes from text prompts, a styling agent for custom map designs, and a Model Context Protocol (MCP) server that links AI assistants to Maps documentation. Additional tools such as Grounding Lite and Contextual View provide low‑code visual answers to location queries. The enhancements also extend to consumer experiences, with hands‑free Gemini navigation and region‑specific alerts in select markets.Weiterlesen

OpenAI’s AI‑video platform Sora has surged to the No. 1 spot in the Google Play Store’s free‑app rankings, overtaking its sister app ChatGPT. The Android version logged 470,000 downloads on its first day, far exceeding the initial iOS launch numbers. New creation tools, cameo features, basic editing, and expanded social options are rolling out, while payment plans now let users purchase extra video generations for $4. Pro users gain longer video limits and storyboarding capabilities. OpenAI’s head of Sora, Bill Peebles, outlined these updates and hinted at further professional‑grade enhancements.Weiterlesen

AI sales startup 1Mind announced a $30 million Series A round led by Battery Ventures, bringing its total funding to $40 million. Co‑founder Amanda Kahlow, formerly of 6Sense, said the company’s AI agent, named Mindy, focuses on inbound sales and can handle tasks from web‑site assistance to closing enterprise deals. Mindy, built on a mix of large‑language models including OpenAI and Google Gemini, uses deterministic AI to limit hallucinations and is designed to say when it does not know an answer. More than 30 companies, including HubSpot, LinkedIn and New Relic, are already using Mindy, and the startup plans to grow its team and product capabilities.Weiterlesen

AI developers are under increasing legal pressure as creators allege that large language model and image generator firms have used copyrighted works without permission to train their systems. More than thirty lawsuits have been filed against firms such as OpenAI, Google, Anthropic, and Meta, while the industry pushes for a fair‑use exemption to keep development costs low. Courts have delivered mixed rulings, with some judges deeming the use "exceedingly transformative" and others allowing settlements. The dispute highlights a clash between the need for rapid AI innovation and the protection of creators’ rights.Weiterlesen

Apple is reportedly developing an artificial‑intelligence health chatbot that would be integrated into a new Health+ subscription service. The chatbot is intended to help users interpret and manage their health data, positioning Apple against competitors that already bundle AI assistants with fitness wearables. While the feature could enhance Apple’s health ecosystem, analysts warn it must deliver tangible value beyond existing tools or risk being seen as unnecessary bloat.Weiterlesen

Enterprises are moving beyond generative AI that merely creates content toward agentic AI systems that act autonomously. While 78% of companies now use generative AI, many see little impact on productivity or revenue. Agentic AI promises to close that gap by turning insights into actions, embedding trust through audit logs and data lineage, linking fragmented tools, and operating continuously without user prompts. Early adopters report faster reporting cycles, reduced compliance costs, and measurable productivity gains, suggesting a shift from surface‑level AI tools to solutions that deliver tangible bottom‑line results.Weiterlesen

Amazon Prime’s biblical drama “House of David” dramatically increased its use of generative AI in its second season, employing roughly four times as many AI‑generated shots as the first. Showrunner Jon Erwin described the approach as a cost‑effective way to create large‑scale battle scenes and expansive landscapes. The production combined image generators, up‑resolution tools, and video generators from vendors such as Runway, Luma, Google and Adobe, using “10 to 15 core tools.” While some critics decried the visual quality, industry labor groups noted the cautious adoption of AI, emphasizing consent and fair compensation for performers.Weiterlesen