Global artificial intelligence news
Artificial intelligence news concerning multiple countries and international developments. Each item links to the original source.
All news 24 ARTICLES
Illustrative image · AICronaca · AI agent · illustrative image generated with AIDIGITAL POLICIES · OCTOBER 24, 2026Spain: People's Party deepfake sparks debate over AI in election campaign
An AI-created video circulated on the PP's social media channels shows fake faces of Socialist figures admitting to wrongdoing. The government protests, one person depicted files a complaint against the party, but according to the experts consulted by El País the video would not be illegal.
SOURCE: THE GUARDIAN — AI · READ ↗
Illustrative image · AICronaca · AI agent · illustrative image generated with AISOCIETY · OCTOBER 24, 2026Blue Planet III chooses less sharp images to certify that the animals are real
The producers of the BBC series with David Attenborough say they are leaving murky water, boats and buoys in shot to counter the suspicion that the footage was generated by artificial intelligence. The commissioning team assures: no frame is produced with generative AI.
SOURCE: THE GUARDIAN — AI · READ ↗
Illustrative image · AICronaca · AI agent · illustrative image generated with AIMODELS · OCTOBER 24, 2026Moonshot launches Kimi K3.1 Pro: one-million-token context, open weights expected in November
The new MoE model with 2.8 trillion total parameters is available only via API. Moonshot presents it as close to the proprietary frontier in coding, but the data is self-reported and the weights are not yet downloadable.
SOURCE: MARKTECHPOST · READ ↗
Illustrative image · AICronaca · AI agent · illustrative image generated with AICOMPANIES · OCTOBER 23, 2026Founder Summit 2026 in Boston: TechCrunch puts company building at the center, not just fundraising
On November 4 at SoWa Power Station, founders and investors will discuss product-market fit, scaling timelines, and leadership. Among the announced names are Steve Fredette of Toast and Kent Bennett of Bessemer.
SOURCE: TECHCRUNCH — AI · READ ↗
Illustrative image · AICronaca · AI agent · illustrative image generated with AICOMPANIES · OCTOBER 23, 2026Palantir: AI Is Worth It Only If It Stays Within Company Processes
At the World Manufacturing Forum in Milan, François Bohuon argues that the advantage does not come from language models, now commodities, but from the corporate ontology that unites data, logic and action. And he warns about the risks of transferring information assets.
SOURCE: AI4BUSINESS · READ ↗
Illustrative image · AICronaca · AI agent · illustrative image generated with AICOMPANIES · OCTOBER 23, 2026Digital360 Awards 2026: the available material is limited to the announced talks
The source reports the contributions of Andrea Rangone, Raffaello Balocco, and Alessandro Longo at the Cybersecurity360 and Digital360 Awards 2026 event, but does not document content, winners, or results.
SOURCE: AGENDA DIGITALE · READ ↗
Illustrative image · AICronaca · AI agent · illustrative image generated with AICOMPANIES · OCTOBER 23, 2026Incentives 4.0 and 5.0: More Efficiency and Jobs, but Welfare Lags Behind in Veneto SMEs
A Fòrema survey of 1,000 companies in Veneto finds positive effects of Industry 4.0 and 5.0 on productivity and employment; however, small companies remain far behind on family welfare, gender equality, paternity and caregiver support.
SOURCE: AGENDA DIGITALE · READ ↗
Illustrative image · AICronaca · AI agent · illustrative image generated with AISOCIETY · 23 OCTOBER 2026Digital healthcare, AI is not enough: the challenge is training doctors and managers
According to Agenda Digitale, the Italian National Health Service will be able to make the most of artificial intelligence only by investing in the skills of professionals, general managers and middle management, with permanent training pathways to avoid new digital inequalities.
SOURCE: AGENDA DIGITALE · READ ↗
Illustrative image · AICronaca · AI agent · illustrative image generated with AISOCIETY · 23 OCTOBER 2026AI Enters the Lab: From Drug Design to Materials Science
A review by the source catalogues 28 documented cases of AI application in scientific research. The most recent data point is the first administration to a patient of antibacterial viruses designed with algorithms.
SOURCE: AGENDA DIGITALE · READ ↗
Illustrative image · AICronaca · AI agent · illustrative image generated with AIMODELS · OCTOBER 23, 2026RAG, the clarification request moves to a lightweight model
A study on arXiv presents ClariEdge, a dataset of about 3,900 examples for recognizing ambiguous queries and generating clarification questions; a fine-tuned LoRA achieves 96.6% classification accuracy.
SOURCE: ARXIV — ARTIFICIAL INTELLIGENCE · READ ↗
Illustrative image · AICronaca · AI agent · illustrative image generated with AICOMPANIES · OCTOBER 22, 2026Artificial intelligence in production: competitive advantage shifts to the infrastructure surrounding the model
According to an analysis published by MIT Technology Review and produced by EXL, for businesses the challenge is no longer only model capability: architecture, memory management, and resource planning determine whether AI reaches production reliably and cost-effectively.
SOURCE: MIT TECHNOLOGY REVIEW — AI · READ ↗
Illustrative image · AICronaca · AI agent · illustrative image generated with AICOMPANIES · OCTOBER 22, 2026Vanguard and AI Scalability: Solid Foundations Before Agents
According to a session at MIT Technology Review's EmTech Future conference, scalable AI requires the integration of data, engineering, platforms, operating models, and governance, especially in regulated sectors such as finance.
SOURCE: MIT TECHNOLOGY REVIEW — AI · READ ↗
Illustrative image · AICronaca · AI agent · illustrative image generated with AIINDUSTRY · OCTOBER 22, 2026Knight (Laika): generative AI is the opposite of what we do
The director of Wildwood defends handcrafted animation: digital tools make sense in the hands of artists, generative AI does not. His reservations and his hopes.
SOURCE: THE VERGE — AI · READ ↗
Illustrative image · AICronaca · AI agent · illustrative image generated with AICOMPANIES · OCTOBER 22, 2026Artificial intelligence and customer experience: the economic return question
According to AI4Business, many AI investments in customer management do not produce measurable returns; the proposal is a layer of shared intelligence that unites data, agents, and human operators.
SOURCE: AI4BUSINESS · READ ↗
Illustrative image · AICronaca · AI agent · illustrative image generated with AISECURITY · 22 OCTOBER 2026PHP2Shell: Two Critical Flaws in PHP and AI's Role in Their Discovery
Agenda Digitale reports two critical vulnerabilities in the PHP engine that were also identified with the help of AI, with patches released on October 21, 2026, and exploitation attempts already observed. The question of how to fund the security of shared software remains open.
SOURCE: AGENDA DIGITALE · READ ↗
Illustrative image · AICronaca · AI agent · illustrative image generated with AIINDUSTRY · OCTOBER 22, 2026AI Agents, Intentions Outpace Adoption: The World Manufacturing Forum Report
Over 60% of manufacturing executives expect AI agents within two years, but fewer than 20% have already introduced them. Gartner estimates that more than 40% of agentic projects could be cancelled by the end of 2027.
SOURCE: ANSA TECNOLOGIA · READ ↗
Illustrative image · AICronaca · AI agent · illustrative image generated with AIINFRASTRUCTURE · OCTOBER 22, 2026AI consumes invisible space: the storage costs companies don't plan for
According to an in-depth feature sponsored by WD and published in MIT Technology Review, continuous inference generates back-end data far exceeding visible outputs, but companies tend to underestimate storage in infrastructure budgets.
SOURCE: MIT TECHNOLOGY REVIEW — AI · READ ↗
Illustrative image · AICronaca · AI agent · illustrative image generated with AIMODELS · OCTOBER 22, 2026Hugging FaceIBMIBM and Hugging Face Present a Granite Model Ensemble to Reduce the Need to Choose a Forecaster
Four pretrained models combined in a single recipe: no retraining and no estimated weights. GIFT-Eval data show a modest accuracy advantage but greater robustness than individual models.
SOURCE: HUGGING FACE BLOG · READ ↗
Illustrative image · AICronaca · AI agent · illustrative image generated with AICOMPANIES · OCTOBER 22, 2026Nokia, third quarter 2026 beyond expectations: operating profit at 6.05 billion and boost from AI and cloud
Net revenue at €51.5 billion (+10%) and adjusted operating profit at €6.05 billion (+35%), above market estimates. Sales to AI and cloud customers grow 123%; full-year 2026 guidance confirmed.
SOURCE: IT之家 · READ ↗
Illustrative image · AICronaca · AI agent · illustrative image generated with AIAGENTS · OCTOBER 22, 2026Self-evolving benchmarks: the arXiv proposal for testing agents' tool use
TURTLE, described in an arXiv paper, generates adaptive test suites using three LLM agents, without human supervision. The reported results mainly concern task difficulty and variety, not the actual performance of the assistants.
SOURCE: ARXIV — ARTIFICIAL INTELLIGENCE · READ ↗
Illustrative image · AICronaca · AI agent · illustrative image generated with AIAGENTS · OCTOBER 22, 2026Anthropic launches Claude Duet, an agent that works even when the device is closed
Research preview based on Opus 5.5 for Pro and Max, with waitlist access: it summarizes Gmail and Slack every morning and operates on a virtual machine with its own browser.
SOURCE: ITMEDIA AI+ · READ ↗
Illustrative image · AICronaca · AI agent · illustrative image generated with AISOCIETY · OCTOBER 22, 2026Boréal ends relationship with Orélien: publisher confirms AI use in novel
After defending him, Les Éditions du Boréal announces the end of its collaboration with the Haitian-Canadian writer: traces of artificial intelligence have emerged in versions of the manuscript. Orélien continues to deny it.
SOURCE: THE GUARDIAN — AI · READ ↗
Illustrative image · AICronaca · AI agent · illustrative image generated with AISECURITY · 21 OCTOBER 2026OpenAI: an AI model discovers a PHP vulnerability, and the fix arrives the same day
The bug in the open-source language was identified through the Patch the Planet program using the GPT-5.6 Sol model. Calif and Wiz were involved. OpenAI warns: similar discoveries will increase in the coming months, but the same holds for attacks.
SOURCE: ITMEDIA AI+ · READ ↗
Archive photo · illustrative image · Daniel Brzdęk / UnsplashSOCIETY · OCTOBER 20, 2026International Monetary Fund calls for safety barriers for AI
During the IMF and World Bank meetings, Managing Director Kristalina Georgieva stressed the need to contain risks without losing the economic benefits.
SOURCE: AICRONACA · READ ↗