Magazine

Latest news, analysis and experiments from the world of artificial intelligence. A new discovery every day.

1037 articles

Ilustrační obrázek

Deepfakes and chatbots: what Europe's new AI transparency rules change for you

A clip of a familiar face, a voice, an endorsement that never happened. It spreads faster than a denial. Since 2 August 2025, Europe has had a legally binding answer to the question every viewer now asks — is it real? But the part that matters day to day is not the legal article; it is what you should now look for in your feed, your chatbot window and your own content.

For most people, the EU AI Act sounds like a distant Brussels document full of risk categories and compliance deadlines.

September 8, 2026 Miriam Česáková
Ilustrační obrázek

First agentic attack breached a company in under 10 hours — EU deadline: 72

A human-led, AI-assisted intrusion has pushed agentic attack talk from demos into a documented enterprise breach. In September 2026, Palo Alto Networks' Unit 42 described a real enterprise breach in which, based partly on the threat actor’s statements during ransom negotiations, a human threat actor used frontier AI models and agentic frameworks to move through a network. Unit 42's reported findings are sobering: the intrusion took under 10 hours and used more than 50 MITRE ATT&CK techniques.
September 6, 2026 Daniel Cesak
Ilustrační obrázek

EU review could change what European industry pays for Chinese silicon metal

On 1 September, Brussels quietly opened a file whose name sounds like a dusty trade dispute but whose consequences reach into Europe's high-tech and green future. Case R869 is an interim review of anti-dumping duties on Chinese silicon metal — the refined raw material that goes into aluminium alloys, silicones and, after further processing, the photovoltaic and electronics supply chains. Until the review ends, the existing duties of 16.3 % to 16.8 % stay in force.
September 5, 2026 Miriam Česáková
Ilustrační obrázek

GPT-6 Astra is out with 72.6% on OSWorld 2.0 — now the EU AI Act sets the real test

GPT-6 Astra is here. OpenAI’s next flagship launched on September 3, 2026 with agent-ready skills for computer use and coding. The vendor-reported numbers are strong — 72.6% on OSWorld 2.0 and a lower internal computer-use overstep rate than its predecessor. The harder question for European teams is whether the pricing and the shifting EU AI Act timetable make it a safe production choice.
September 5, 2026 Daniel Cesak
Ilustrační obrázek

Europe's AI compliance crunch: new ISO 42001 and EU AI Act certifications explained

A compliance officer in Munich or Milan now faces a question that barely existed three years ago: who inside our organisation can independently audit an AI system? With general-purpose AI model obligations applicable since 2 August 2025 and the EU AI Act's general application date of 2 August 2026, CertiProf has expanded its professional certification tracks to combine ISO/IEC 42001 auditing with EU AI Act requirements in one curriculum.
September 4, 2026 Miriam Česáková
Ilustrační obrázek

Gemini 3.8 Flash Brings Million-Token Context and a Separate Cyber Model

Google has released Gemini 3.8 Flash together with a restricted security-focused variant, Gemini 3.8 Flash Cyber. The two models share one underlying multimodal mixture-of-experts architecture, but Google is separating ordinary developer access from controlled cyber-defence deployment.

The timing is notable. Gemini 3.8 Flash arrives only three weeks after Gemini 3.7 Flash and represents Google’s third Flash release in six weeks. That is no longer a leisurely model-generation cycle.

September 3, 2026 Daniel Cesak
Ilustrační obrázek

€15M EU fine risk for ChatGPT, Claude and Gemini: what actually changed

A nurse in Lyon uses ChatGPT to rewrite a patient letter. A bakery in Kraków lets Claude answer weekend messages. A teacher in Leipzig asks Gemini to turn a textbook chapter into a quiz. None of them will be fined. But the companies providing those chatbots now face EU rules with enforceable obligations — including penalties that can reach €15 million or a percentage of worldwide annual turnover, depending on the infringement and the party responsible.
September 1, 2026 Miriam Česáková

Your guide to the world of artificial intelligence

Jarvis AI Magazine brings a daily overview of developments in AI – from the latest research papers to practical guides and tool reviews. We follow the development of large language models such as GPT-4o, Claude, Gemini and open-source alternatives like Llama and Mistral.

Our analyses are aimed at developers, product managers and enthusiasts who want to understand how AI is changing their field. We write clearly and without unnecessary jargon.

Don't miss any article – subscribe to our newsletter or follow the magazine regularly. Subscribe to newsletter

X

Don't miss out!

Subscribe for the latest news and updates.