European enterprises are deploying AI agents faster than they establish governance frameworks, resulting in security incidents involving non-human identities.
In 2026, AI funding will only be granted to projects with demonstrable development risk, with the EU AI Act—which is being phased in starting 2025—serving as the baseline for eligible solutions.
HarnessX automates the assembly and adaptation of agent harnesses from execution traces, achieving an average +14.5% performance improvement without model scaling.
A new benchmark enables identification of the exact point where medical AI models produce hallucinations and enables targeted countermeasures through trace-supervised fine-tuning.
A trainable classifier predicts with a 0.7 Macro-F1-Score based on early hidden states whether activation steering will succeed without requiring complete generations.
Language models are evolving from chatbots with simple next-token prediction into Digital Colleagues with working memory, persistent workspaces, reusable skills, and reliable problem-solving.
A security vulnerability reported by Amazon to the White House triggered an AI model release freeze, exposing deep structural tensions between private AI makers, the US government, and national export control interests.
