I'm on a hunger strike outside DeepMind's office in London. Here's what I fear most about AI .
Trazzi fears AI models like GPT-5 and Gemini 2.5 Pro are nearing dangerous capabilities. ... Yet, the real danger comes when AI can automate its own ...
Read the original at Business Insider ↗
- Topics
- AI Safety & Existential RiskAI Agents & Automation
- Mentions
- GoogleGPTGeminiGoogle DeepMindOpenAI
- Tracked via
- “dangerous AI”