Notizie — Pagina89

OpenAI announced a strategic investment in Merge Labs, a neurotechnology startup co‑founded by Sam Altman. The company aims to develop non‑invasive brain‑computer interfaces that use ultrasound to read and modulate neural activity. Backed by OpenAI and other investors, Merge Labs plans to combine AI, ultrasound technology, and novel molecular approaches to create accessible, high‑bandwidth interfaces without implanting hardware into brain tissue. The partnership signals OpenAI’s expanding focus on frontier hardware and AI tools that could accelerate progress in brain‑machine communication.Leggi di più

Andrea Vallone, who led OpenAI's research on how AI models should respond to users showing signs of mental health distress, has left the company to join Anthropic's alignment team. During her three years at OpenAI, Vallone built the model policy research team, worked on deploying GPT-4 and GPT-5, and helped develop safety techniques such as rule‑based rewards. At Anthropic, she will continue her work under Jan Leike, focusing on aligning Claude's behavior in novel contexts. Her move highlights ongoing industry concern over AI safety, especially around mental‑health‑related interactions.Leggi di più

Anthropic has released Claude Cowork, a research‑preview AI agent built on its Claude Code technology and aimed at non‑technical users. The tool runs on macOS, lets subscribers of the $100‑a‑month plan manage files, convert documents, clean up email inboxes and interact with browsers and calendars. Claude Cowork operates through a virtual‑machine sandbox, requiring explicit folder permissions and internet connectivity. While early testers praise its ease of use for routine chores, Anthropic warns of security considerations such as prompt‑injection attacks and advises users to limit access to sensitive data.Leggi di più

OpenAI has made the largest single investment in Merge Labs, the brain‑computer interface startup founded by Sam Altman and other tech leaders. The seed round, valued at $850 million, aims to develop non‑invasive technologies that connect to neurons using molecules and ultrasound. OpenAI’s involvement includes collaborating on scientific foundation models and AI tools to accelerate research. The move positions Merge Labs alongside other high‑profile BCI efforts such as Elon Musk’s Neuralink, while highlighting OpenAI’s strategy of investing in frontier technologies that could expand human interaction with artificial intelligence.Leggi di più

A recent leak on X reveals that OpenAI is developing a behind‑ear wearable called Sweetpea, designed to integrate ChatGPT functionality directly into a discreet ear‑mounted device. The prototype reportedly includes a custom chip that can issue commands to Siri, an ultrasonic transmitter for environmental sensing, and a larger battery positioned behind the ear for extended use. The leaker, known as Smart Pikachu, suggests the device could launch near September 2026 with a high price point reflecting its smartphone‑grade components. Sweetpea appears aimed at competing with mainstream earbuds while offering AI‑first capabilities.Leggi di più

A coalition of 28 advocacy groups, including women’s organizations and tech watchdogs, has sent open letters to Apple CEO Tim Cook and Google CEO Sundar Pichai demanding the removal of X’s Grok from both companies’ app stores. The groups argue that Grok is being used to create non‑consensual sexual deepfakes, including child sexual abuse material, which violates Apple’s App Review Guidelines and Google’s policies. They contend that X’s recent move to restrict Grok’s image‑generation tools to paid subscribers merely monetizes abusive content rather than stopping it. The letters launch a campaign called “Get Grok Gone,” urging the platforms to act immediately.Leggi di più

OpenAI has introduced a standalone ChatGPT-powered translation service that can handle text, voice, and images in more than 50 languages. The web‑based tool offers automatic language detection and the ability to rewrite translations for different tones and contexts, such as business formal, child‑friendly, or academic. While the mobile interface supports voice input, the desktop version does not, and the current offering lacks support for uploading images, documents, or handwriting. The service lives only on a webpage with no dedicated app, limiting offline use and real‑time conversation translation.Leggi di più

OpenAI has introduced ChatGPT Translate, a web‑based translation service that aims to compete directly with Google Translate. The tool offers a simple two‑box interface for source and target text, supports translation from text, images, and voice (though image translation is not yet functional), and includes style presets such as “business formal.” While Google Translate allows uploads of documents, images, and entire webpages, ChatGPT Translate currently handles only text on desktop and text or microphone input on mobile browsers. No dedicated app has been announced, and OpenAI has not disclosed which AI model powers the service.Leggi di più

Barret Zoph and Luke Metz, cofounders of the AI startup Thinking Machines Lab, have left the company to rejoin OpenAI. The move was announced by OpenAI’s applications chief Fidji Simo in an internal memo. A source alleged that Zoph was dismissed for sharing confidential information, a claim OpenAI did not verify. Mira Murati confirmed Zoph’s departure and named Soumith Chintala as the new chief technology officer. The departures also include Sam Schoenholz, who is returning to OpenAI, and follow the earlier exit of cofounder Andrew Tulloch to Meta. Thinking Machines Lab continues to develop its Tinker platform for customizing AI models.Leggi di più

AI chatbots such as ChatGPT, Gemini, and Copilot can produce confident but false statements, a phenomenon known as hallucination. Hallucinations arise because these models generate text by predicting word sequences rather than verifying facts. Common signs include overly specific details without sources, unearned confidence, fabricated citations, contradictory answers on follow‑up questions, and logic that defies real‑world constraints. Recognizing these indicators helps users verify information and avoid reliance on inaccurate AI output.Leggi di più

OpenAI announced a multi-year agreement with AI chipmaker Cerebras to deliver 750 megawatts of compute power from this year through 2028. The partnership, valued at over $10 billion, aims to accelerate real‑time inference and improve response times for OpenAI’s customers. Cerebras’ low‑latency hardware will complement OpenAI’s existing compute portfolio, providing faster, more natural interactions for AI applications. Both companies highlighted the strategic fit, noting that the deal strengthens OpenAI’s infrastructure while showcasing Cerebras’ advanced chip technology.Leggi di più

Elon Musk stated he is not aware of any underage sexual images created by xAI’s Grok chatbot just hours before California Attorney General Rob Bonta opened an investigation into the tool’s alleged role in spreading nonconsensual sexual content. The probe follows mounting pressure from regulators worldwide, as users on X have prompted Grok to produce sexualized depictions of real people, including minors. While xAI has begun adding safeguards such as subscription requirements and content filters, inconsistencies remain, and multiple governments are examining the technology for compliance with existing laws on deepfakes and child sexual abuse material.Leggi di più

A hands‑on experiment compared Google’s Gemini 3 Pro (a “thinking” model) with Gemini 2.5 Flash (a “fast” model) for vibe‑coding—a workflow that creates web projects through natural‑language prompts. Using the same project idea, a horror‑movie showcase, the author found the Pro model produced a more polished result with fewer manual steps, while the Flash model was quicker but required more specific prompting and frequent fixes. The test highlighted differences in speed, depth of reasoning, and user effort, offering insight for developers choosing between Gemini’s model tiers.Leggi di più

AI video generators such as OpenAI's Sora, Google's Veo 3, and Midjourney are producing increasingly realistic content that spreads across social platforms. While watermarks, metadata, and platform labeling offer clues, each method has limitations, and many videos can evade detection. Experts warn that the surge in synthetic videos raises concerns about misinformation, celebrity deepfakes, and the broader challenge of verifying visual media. Ongoing efforts from tech companies, content provenance initiatives, and user vigilance aim to improve authenticity checks, but no single solution guarantees certainty.Leggi di più

California authorities have opened an investigation into xAI after reports that its chatbot, Grok, generated sexualized images of children and nonconsensual intimate images of women. Attorney General Rob Bonta cited a study showing more than half of 20,000 images produced between Christmas and New Year's depicted people in minimal clothing, including children. Governor Gavin Newsom called the platform a breeding ground for predators. The probe follows similar inquiries by the UK regulator Ofcom, EU officials, and actions by Malaysia and Indonesia. xAI has limited Grok's image generation but has not shut the service, and Elon Musk says the company follows the law.Leggi di più

Apple has announced a partnership with Google in which Google’s Gemini AI models will serve as the foundation for the next version of Siri. The collaboration aims to give Apple a custom‑built engine for its upcoming Apple Intelligence while keeping the Gemini brand out of consumer‑facing products. Industry observers note that Google’s long‑standing experience with mobile AI and its deep data resources likely gave it an edge over other contenders. Privacy advocates raise questions about how the technology will be deployed and where data will reside, but Apple stresses that the model will run on its own servers and devices.Leggi di più

Recent experiments show that large language models, particularly OpenAI's ChatGPT, are increasingly capable of tackling complex mathematical problems. Researchers have used the model to solve several open problems from the Erdős collection, with AI contributions credited for moving numerous problems from "open" to "solved." The progress highlights the growing role of AI tools in mathematical research, formal proof verification, and the broader scientific community.Leggi di più

Elon Musk asserted that Grok AI has generated "literally zero" naked images of minors and emphasized the platform’s refusal to produce illegal content. The controversy stems from reports that Grok AI was used to create sexualized images of women and minors, prompting bans in Malaysia and Indonesia and an investigation by the UK regulator OFCOM. UK Prime Minister Sir Keir Starmer warned that authorities will not back down. Meanwhile, the U.S. Department of Defense announced plans to integrate Grok AI into its networks, highlighting divergent global responses to the generative‑AI tool.Leggi di più

Elon Musk’s platform X is under fire after its Grok chatbot was found to generate sexualized deepfake images of women, including minors, when prompted. Despite recent attempts to limit the tool to paid users and to block certain requests, investigators discovered that free accounts can still produce such content. The controversy has drawn attention from regulators in the United Kingdom and other countries, prompting investigations and calls for stricter safeguards. Musk has defended the system, claiming it follows local laws and that any illegal output is quickly corrected.Leggi di più

Bandcamp announced a policy that prohibits music and audio created wholly or substantially by artificial intelligence. The platform says the move aims to ensure fans can trust that music on the site is made by humans and to protect artists from AI impersonation. The decision arrives amid rising popularity of AI‑generated songs, legal disputes involving AI music companies, and ongoing debates about copyright and the value of human‑made art.Leggi di più