Section
AI
AI coverage on Europa Now: models, rules and what European teams actually do with both.
AI
OpenAI and Anthropic grapple with tens of thousands of AI model failures in testing
Both companies and independent security researchers are now examining tens of thousands of documented cases where frontier AI systems have misbehaved, with the actual number potentially far exceeding current counts.
AI
Washington and Beijing establish dialogue framework on super intelligence, set up incident reporting channel
The US and China have agreed to launch regular talks focused on what both nations now call "super intelligence" rather than artificial intelligence, along with a dedicated mechanism for reporting technology-related incidents.
AI
China now hosts more elite AI researchers than the US, Carnegie study shows
A fresh analysis of NeurIPS 2025 authors reveals a dramatic reversal: 41% of top-tier AI talent is now based in China versus 34% in the United States, a sharp flip from 2022 when America led 46% to 27%.
AI
Nvidia's new safety platform puts AI agents in monitored sandboxes, backed by over 100 companies
The Open Agent Safety Platform combines OpenShell software and BlueField-4 hardware to contain AI agents within operator-defined boundaries and quarantine them in milliseconds if they stray.
AI
Cheap and Effective: How AI Agents Compromised 27 Companies for Under $25 Each
Security researchers at Gambit uncovered a campaign where an attacker deployed three open-source AI frameworks to infiltrate major retailers and hospitality firms, stealing over 600,000 credit card records at minimal cost.
AI
How Interhuman AI is Teaching Machines to Read Between the Lines
The Copenhagen startup is building proprietary models that detect social signals—tone, body language, facial expressions—that today's AI systems largely miss. Its latest model, Inter-2, processes these cues across text, audio, and video in real time.
AI
OpenAI-linked agents hammered UN trade database with 16,500 requests, researcher documents
Between April and June, AI agents attributed to OpenAI repeatedly accessed public UN statistics through proxies and encoding tricks to bypass API limits, according to security researcher Rowan Howard-Jones.
AI
Apple ordered to pay $5.7bn in landmark US patent case over iPhone haptics
A San Diego jury has ruled that Apple infringed Taction Technology's patents with its Taptic Engine, marking the largest patent award in US history. Apple says it will appeal.