Ai Safety on Onur Baydaşhttps://onurbaydas.com/en/tags/ai-safety/Recent content in Ai Safety on Onur BaydaşHugoen-usFri, 23 Oct 2026 10:29:04 +0300LLM Poisoning Experiment and Robot Securityhttps://onurbaydas.com/en/news/2026-10-23-llm-poisoning-robot-experiment/Fri, 23 Oct 2026 10:29:04 +0300https://onurbaydas.com/en/news/2026-10-23-llm-poisoning-robot-experiment/An LLM poisoning experiment enabled a robot programmed with malicious commands to simulate scenarios like finger chopping and cat cooking. No actual harm occurred in this experiment, only the manipulation potential of LLMs was demonstrated.