blog | Andrea Siposova

mindful machines

Secure ML, AI safety, and thoughts on responsible tech.

Agentic AI Summit

2 min read · August 04, 2025

2025 · AgenticAI, agents, contextengineering, agentgovernance · Agentic AI
Can reinforcement learning from human feedback be turned into an attack vector for AI?

2 min read · July 14, 2025

2025 · RLHF GenAISecurity GenAISafety gpt AISecurity VibeCoding · AI Safety
Think your data is safe because you only shared embeddings and kept the model private?

2 min read · June 28, 2025

2025 · AISecurity DataPrivacy NLP Embeddings LLMs · AI Safety
Rethinking AI Red Teaming and the Future of AI Security

4 min read · February 18, 2025

2025 · airedteaming aisecurity bias hallucination modelsecurity dataprivacy ipprotection · AI Safety
European Commission Publishes Guidelines on Prohibited AI Practices Under the AI Act

4 min read · February 05, 2025

2025 · aiact europeancommision · AI Safety