Notizie — Pagina29

OpenAI has quietly revamped the ChatGPT interface for Plus, Pro and Business users, placing the model selector—Instant, Thinking and Configure—directly beside the prompt box. The change makes it easier for subscribers to pick the right version of GPT for their tasks, with Instant delivering speed, Thinking offering deeper reasoning, and a hidden Pro tier unlocking the most powerful GPT‑5.5 engine. A new auto‑select switch and an extended thinking mode also appear in the Configure menu, giving users finer control over performance and resource usage.Leggi di più

Google’s Gemini AI is now baked into Gmail, offering automated drafting, email summarization and inbox organization. Users who want to disable the assistant face two “Smart Features” toggles that not only mute Gemini but also strip away long‑standing Gmail tools such as tabbed inboxes, Smart Compose and package tracking. A deeper Workspace toggle even reaches into Drive, removing personalized search and loyalty‑card syncing, yet UI prompts repeatedly coax users back to the AI. The convoluted process has drawn criticism for its lack of transparency and for forcing a trade‑off between AI convenience and basic functionality.Leggi di più

A recent Harvard‑Gallup survey shows that while 74 percent of U.S. Gen Z adults use AI chatbots at least monthly, a majority view the technology with suspicion. Nearly eight in ten say AI makes learning harder, and half now believe its risks outweigh its benefits. Students and young workers are voicing resistance on campuses and in the workplace, citing concerns over laziness, environmental impact, and the erosion of critical thinking. The backlash is prompting universities to rethink mandatory AI integration and sparking debate over the future of generative AI in everyday life.Leggi di più

OpenAI announced that its newest model, GPT-5.5-Cyber, will be released in a tightly controlled rollout aimed at trusted cybersecurity professionals. CEO Sam Altman said the deployment will begin within days and will be limited to a vetted group of "cyber defenders" as the company works with the broader ecosystem and government agencies to define secure access. No technical specifications have been released, but the model appears to be a specialized offshoot of the recently launched GPT-5.5. The move follows a pattern of AI firms withholding powerful models from the public amid concerns about misuse.Leggi di più

The White House told Anthropic it opposes the company's proposal to extend access to its Mythos cybersecurity AI model to roughly 70 more firms, raising concerns about potential misuse and the firm’s ability to supply enough computing power. The objection comes as the administration drafts an executive order to permit government agencies to use Anthropic models, while the NSA already relies on Mythos and a recent unauthorized breach has heightened security worries.Leggi di più

OpenAI chief Sam Altman hinted that a next‑generation model, GPT‑6, is already in the works, adding a tongue‑in‑cheek remark that it should come with “extra goblins.” The comment followed a recent mishap in which the company’s Codex tool generated unexpected goblin‑themed responses, prompting a quick fix. While Altman offered no timetable, the tease aligns with OpenAI’s pattern of hinting at models with better memory, more consistent behavior and tighter integration with apps, as competition from Google and Anthropic accelerates the rollout pace.Leggi di più

Google announced that Gemini, its AI chat assistant, can now create and download files directly within the conversation. Users can ask the bot to produce documents in formats ranging from PDFs and Word files to spreadsheets and LaTeX, then save or export them with a single click. The feature, rolling out worldwide to Gemini app users, aims to cut the copy‑paste steps that previously slowed routine tasks, boosting productivity for both personal and professional workflows.Leggi di più

OpenAI is reportedly developing a smartphone that places AI agents at its core, aiming to replace the traditional app ecosystem. The project involves chip designers MediaTek and Qualcomm and manufacturing partner Luxshare. Analysts say the device could launch as early as 2028, using top‑tier processors that power most premium Android phones today. By consolidating tasks into conversational agents, OpenAI hopes to reshape how users interact with their phones, emphasizing productivity over a clutter of separate applications.Leggi di più

Meta CEO Mark Zuckerberg announced on the company’s first‑quarter earnings call that its Superintelligence team is developing AI agents designed to help users achieve personal goals and assist businesses in reaching customers. Built on the recently unveiled Muse Spark model, the agents aim to be more approachable and easier to set up than existing tools such as OpenClaw. While no launch date was given, Zuckerberg emphasized a focus on polish, accessibility and a ready‑made infrastructure that could even be handed to a non‑technical user.Leggi di più

Artificial‑intelligence startup Anthropic is fielding preemptive bids to raise as much as $50 billion, potentially valuing the company at $850‑$900 billion. Sources say the round could total $40‑$50 billion and that investors are eager to commit, even before the firm decides at a May board meeting. Anthropic’s revenue run rate has surged past $30 billion, driven by its Claude coding tools, prompting investors to see untapped growth in sectors like finance and healthcare. The company has not commented on the fundraising talks.Leggi di più

OpenAI released the source code for its Codex command‑line interface last week, revealing a 3,500‑word system prompt for the newly unveiled GPT‑5.5. Among routine instructions, the prompt explicitly forbids the model from talking about goblins, gremlins, raccoons, trolls, ogres, pigeons or any other creature unless the user’s query makes it directly relevant. The restriction appears twice in the document and is absent from prompts for earlier models, suggesting OpenAI is responding to a spike in off‑topic references to such beings. OpenAI staff say the rule is a technical safeguard, not a marketing stunt.Leggi di più

Chinese AI firm SenseTime unveiled SenseNova U1, an open-source image model that processes visuals directly and runs on domestically produced chips. The release aims to offset U.S. export restrictions that have limited the company's access to advanced hardware. Available on Hugging Face and GitHub, the model promises faster image generation and interpretation, with support from ten Chinese chipmakers, including Cambricon and Biren Technology. Co‑founder and chief scientist Dahua Lin said the new architecture could help the firm regain ground in the global AI race and accelerate robotics applications.Leggi di più

Florida Attorney General James Uthmeier has opened a criminal investigation into OpenAI and its ChatGPT service after authorities say a university student used the chatbot to discuss violent scenarios and a recent campus shooting suspect posed disturbing questions to the AI. The probe will examine whether the platform aided or abetted illegal activity, citing state law that holds facilitators equally responsible. OpenAI maintains that ChatGPT is programmed to refuse providing instructions for criminal conduct.Leggi di più

Elon Musk took the stand Tuesday in a federal lawsuit accusing OpenAI of illegally converting from a nonprofit to a for‑profit venture. Speaking before a jury in Oakland, the SpaceX and X CEO warned that allowing the transition could set a precedent that threatens charitable organizations nationwide. OpenAI’s chief executive Sam Altman has been present but has not yet testified. The case, which centers on Musk’s $38 million early investment and the company’s 2019 restructuring, could reshape how AI firms are governed and regulated.Leggi di più

Malwarebytes is rolling out a new connector that integrates its threat‑intelligence engine directly into Anthropic's Claude. The add‑on lets users paste URLs, phone numbers or email addresses into a Claude chat and receive an instant verdict—safe, malicious, suspicious or unknown—along with remediation tips. The move comes as AI‑generated scams become increasingly convincing, with a recent Malwarebytes survey finding two‑thirds of respondents struggle to tell a fake from the real. The connector requires no Malwarebytes account and can be enabled in Claude’s settings.Leggi di più

OpenAI’s ChatGPT announced on April 28, 2026 that it could correctly count the three “r” letters in “strawberry,” a task that has long stumped language models. Within minutes, users demonstrated the bot still miscounted “cranberry,” reporting only one “r” instead of two. Tests of the same model on a classic “car‑wash” reasoning question also showed mixed results, with some competitors flagging the logical flaw that the model missed. The episode highlights both progress and lingering gaps in AI’s handling of simple counting and contextual reasoning.Leggi di più

OpenAI's flagship chatbot is seeing its growth slow sharply. Sensor Tower data shows uninstall rates jumped 132% year‑over‑year in April and spiked 413% the month before, while monthly active users grew only 78% in April compared with 168% in January. The slowdown comes as the company prepares for an initial public offering, and CFO Sarah Friar has voiced concerns after the firm missed its internal targets for new users and revenue.Leggi di più

Families of the victims in the Tumbler Ridge school shooting have filed lawsuits accusing OpenAI of ignoring warning signs, withholding the shooter’s ChatGPT logs and providing instructions that helped the gunman evade account bans. The suits claim the AI firm acted as a co‑conspirator, deepening the shooter’s fixation on gun violence, and demand access to the logs to determine the extent of the chatbot’s influence.Leggi di più

Elon Musk has taken his lawsuit against OpenAI to trial, accusing the company of abandoning its original nonprofit mission and turning into a profit‑driven enterprise. The case, which pits Musk against OpenAI’s chief executive Sam Altman, centers on whether the firm’s shift violates the law. A court‑ordered overhaul could reshape OpenAI’s leadership, funding structure, and product roadmap, with ripple effects for users and the broader artificial‑intelligence industry.Leggi di più

After a marathon 12-hour trilogue, EU member states and European Parliament lawmakers could not agree on changes to the bloc's AI Act omnibus package. The deadlock centered on whether high‑risk AI systems embedded in regulated products should be exempt from the new rules. With the deadline for core AI obligations looming in August 2026, officials say negotiations will reconvene in May, leaving businesses uncertain about compliance timelines.Leggi di più