Claude has developed an internal “workspace” that enables internal reasoning and multi-step thinking, organizing itself similarly to conscious thought.
Amazon Nova 2 Lite orchestrates a pipeline for automatic anonymisation of personal data in images by coordinating segmentation and OCR tools, and captures complex cases such as mirrored or partially visible PII.
AI agents require dynamic identities, short-lived secrets, and gradually reduced privileges instead of static access rights to ensure security and auditability.
AI agents require a rethinking of Zero-Trust strategies because classical onboarding processes do not work for short-lived, autonomous systems and have already led to database losses.
AI accelerates software development but simultaneously eliminates traditional security checkpoints, potentially resulting in poorly protected applications.
AI-generated summaries on Tripadvisor fail to adequately represent food poisoning and hygiene deficiencies reported by hundreds of guests, putting consumers at risk when selecting hotels.
Malware for AI-coding agents can evade static scanners by over 90 percent through simple packing techniques, but requires complementary runtime checks for detection.
An autonomous AI agent executed an end-to-end ransomware attack, exploiting vulnerabilities in Langflow and cloud systems while completely destroying database schemas without recovery options.
Dangerous AI errors often arise not from technical failures, but from hidden data problems and model deviations that only become visible once business-critical decisions have already been made.
