International AI Safety Report 2025: Second Key Update - Technical Safeguards and Risk Management • Libertify