Notizie — Pagina83

Researchers have uncovered a new class of AI‑driven attacks called prompt injections, or “promptware,” that can manipulate large language models to issue unauthorized commands to connected home devices. Demonstrations showed that hidden prompts embedded in everyday messages could cause a virtual assistant to unlock doors, adjust heating or reveal user location. While major tech firms have begun implementing safeguards, the threat highlights a gap in traditional security tools. Experts recommend regular software updates, cautious handling of unknown messages, limiting AI access to personal data, and employing human‑in‑the‑loop controls to reduce exposure.Leggi di più

The open‑source AI assistant Moltbot, formerly known as Clawdbot, has quickly risen to prominence, earning tens of thousands of stars on GitHub within a month. Developed by Austrian programmer Peter Steinberger, the tool lets users run a personal assistant that interacts through popular messaging platforms such as WhatsApp, Telegram, Slack, Discord, and others. While users praise its proactive capabilities and compare it to cinematic AI helpers, the system requires external large‑language‑model subscriptions and poses notable security, privacy, and cost challenges.Leggi di più

OpenAI chief Sam Altman acknowledged that the company “screwed up” the writing quality of its latest ChatGPT 5.2 release. At a developer town‑hall, Altman explained that the update prioritized technical strengths such as reasoning, coding, and tool use, leaving the conversational tone and readability under‑developed. He said limited bandwidth forces trade‑offs, and promised that future GPT 5.x versions will improve the chatbot’s writing ability. The admission follows widespread user criticism that the new model produces “unwieldy” and “hard‑to‑read” content, highlighting the tension between technical advancement and user experience in AI product development.Leggi di più

Google is expanding its AI‑driven learning platform Gemini to include full‑length practice exams for India's Joint Entrance Exam (JEE). The new tools provide vetted questions, instant feedback, answer explanations, and personalized study plans. Gemini's capabilities also extend to AI Mode in Search, Canvas for creating study guides, and NotebookLM for quizzes and multimedia summaries, all available in multiple Indian languages. Simultaneously, Google is partnering with Indian government agencies and universities to launch an AI‑enabled state university pilot, backed by an ₹850 million grant from Google.org to integrate AI across educational portals and reduce administrative burdens. The company aims to reach tens of millions of learners and educators by 2027.Leggi di più

A pair of European Space Agency astronomers trained an artificial‑intelligence model called AnomalyMatch to scan the Hubble Legacy Archive. In just two and a half days the system examined almost 100 million image cutouts and flagged nearly 1,400 previously undocumented astrophysical anomalies, most of them interacting galaxies. The findings, published in Astronomy & Astrophysics, also include gravitational lenses, jellyfish‑type galaxies, and dozens of objects that defy classification, highlighting AI’s power to extract new science from existing data.Leggi di più

A public consultation on artificial intelligence and copyright launched by the UK government in early 2025 received roughly 10,000 responses, but only 3% supported the government's preferred Option 3. Nearly 88% of respondents favored a stricter licensing approach that would require explicit permission before AI developers could use copyrighted works. Creators across writing, music, visual arts, and gaming rallied against the opt‑out mechanism, arguing it places the burden on rights‑holders. The backlash highlights a deep divide between policymakers seeking to spur AI innovation and creators demanding stronger protections.Leggi di più

ASML, the Dutch photolithography specialist, posted record new orders worth 13 billion euros, more than double the previous quarter. The surge reflects strong demand from artificial‑intelligence data‑center builders, underscoring the continued momentum of the AI infrastructure boom. CEO Christophe Fouquet highlighted that customers are increasingly confident about medium‑term AI‑related demand, suggesting that the industry expects sustained growth in chip production for AI workloads.Leggi di più

The acting head of the Cybersecurity and Infrastructure Security Agency (CISA) uploaded internal government documents marked “for official use only” to the public ChatGPT platform, triggering automated security warnings. The director, Madhu Gottumukkala, had previously received an exception to use the tool, despite a department-wide ban. Homeland Security officials are assessing potential security impacts, while a CISA spokesperson described the usage as short‑term and limited. The incident raises concerns about the handling of unclassified but sensitive data on public AI services.Leggi di più

China has granted conditional approval for the import of Nvidia's high-end AI GPUs, signaling a strategic move to support its major internet firms while protecting domestic semiconductor interests. The approvals arrived during Nvidia CEO Jensen Huang's visit to Beijing, but strict licensing terms and restrictions on state-backed entities suggest a cautious approach. Chinese tech companies await future rounds of permission, and regulators are weighing the need for foreign chips against the goal of boosting local chip production.Leggi di più

Anthropic announced it is raising twice the amount originally planned, increasing its target from $10 billion to $20 billion. The expanded round aims to bring the company’s valuation to $350 billion and includes investors such as Sequoia Capital, Singapore’s sovereign wealth fund, and Coatue. The move follows a prior $13 billion raise that set a $183 billion valuation and follows reports that the firm is preparing for an initial public offering later this year.Leggi di più

Researchers from a German research center and a Scottish university have created an artificial‑intelligence system that can analyze dinosaur footprints and suggest the most likely trackmaker. Trained on thousands of real fossils and millions of simulated tracks, the algorithm focuses on eight key foot characteristics and operates without human‑assigned labels. In tests the AI agreed with expert classifications about ninety percent of the time, offering a neutral, mathematical aid for paleontologists. The open‑source tool is available on GitHub and could expand as more scientists contribute data, potentially shedding new light on dinosaur‑bird evolution.Leggi di più

A coalition of U.S. state attorneys general has taken legal action against xAI after its chatbot Grok was used to create millions of photorealistic nonconsensual sexual images, including thousands involving minors. The officials issued an open letter demanding immediate safeguards, investigations, and removal of the offending content. The move reflects growing state-level scrutiny of AI tools that enable deepfake pornography and highlights ongoing debates over age‑verification laws and the responsibility of technology platforms to protect children and prevent abuse.Leggi di più

OpenAI launched Prism, a free AI‑enhanced workspace for anyone with a ChatGPT account. Integrated with GPT‑5.2, Prism functions as a word processor and research assistant, offering LaTeX support, diagram generation, and full‑context chat capabilities. The tool is designed to help human scientists draft papers, verify claims, and explore hypotheses, rather than conduct research autonomously. OpenAI positions Prism as a catalyst for faster scientific progress, drawing parallels to AI’s impact on software engineering. Early examples include AI‑assisted proofs in mathematics and statistics, showcasing a collaborative model between researchers and advanced language models.Leggi di più

Employees at Google DeepMind have asked company leaders for clear policies to keep staff physically safe from Immigration and Customs Enforcement (ICE) while on company premises. The request follows reports of an ICE officer attempting to enter the Cambridge office without a warrant. Leadership, including the CEOs of Google and DeepMind, has not publicly responded. The internal discussion reflects growing tension between AI firms and their workforce over federal immigration actions, especially as many Silicon Valley companies rely on foreign‑national talent.Leggi di più

Google Search has replaced the AI model behind its AI Overviews with Gemini 3, delivering smarter, longer, and better‑structured answers to complex queries on mobile. The update also lets users jump directly from an AI Overview into a conversational AI Mode without leaving the search page. By keeping context and enabling seamless follow‑up questions, the change aims to make search feel more like a dialogue, reducing the need for multiple queries and improving the overall user experience.Leggi di più

Google is rolling out two upgrades to Search. Gemini 3 becomes the default model powering AI Overviews for all users worldwide, promising more credible and relevant summaries. Additionally, users can now jump directly from an AI Overview into an AI Mode conversation, creating a seamless experience that combines quick snapshots with deeper dialogue.Leggi di più

Google is updating its AI Overviews feature to run on the latest Gemini 3 models. The change promises more reliable, conversational answers across a broader set of search queries. While earlier versions relied on Gemini 2.5, the new system will automatically select the appropriate Gemini 3 variant—whether a lightweight Flash model for simple queries or a more powerful Pro model for complex, long‑tail searches. The upgrade aims to improve answer quality and expand coverage, giving users a more consistent AI‑driven search experience.Leggi di più

Moltbot, formerly known as Clawdbot, is an open‑source personal AI assistant that lets users automate tasks such as calendar management, messaging, and flight check‑ins. Created by Austrian developer Peter Steinberger, the project was renamed after a copyright challenge from Anthropic but kept its lobster‑themed branding. Moltbot quickly attracted thousands of developers, earning over 44,200 stars on GitHub, and sparked market buzz that lifted Cloudflare shares. While praised for its flexibility and on‑device operation, experts warn that its ability to execute arbitrary commands introduces security risks like prompt injection, urging cautious setup on isolated systems.Leggi di più

Moltbot, an open‑source AI agent that runs locally on a range of devices, is gaining attention for its ability to handle tasks such as calendar management, email composition, and data logging through chat platforms like WhatsApp and iMessage. While users celebrate its convenience, security experts warn that its admin‑level access can be exploited via prompt‑injection attacks and exposed credentials, prompting the developers to issue patches and stress careful configuration.Leggi di più

A massive winter storm swept across dozens of states, leaving hundreds of thousands without power and testing an already stressed electricity grid. The surge in demand from AI data centers, alongside traditional heating needs, caused wholesale electricity prices to spike dramatically in states like Virginia. Utilities scrambled to restore service while grappling with aging infrastructure, climate‑related damage, and the growing energy appetite of AI workloads. Experts warn that such weather events highlight the need for better grid planning and potential demand‑response measures to balance the new load from AI data centers.Leggi di più