OpenAI CEO Sam Altman is pushing the United States to eliminate pre-approval before launching new AI models. His argument: bureaucracy hinders progress. But without controls, companies can release systems with biases, security flaws, or dangerous capabilities without oversight. History shows that self-regulation fails.
No technical filters: vulnerabilities and biases exposed 🚨
Current AI models exhibit racial and gender biases, as well as serious errors in medical or judicial contexts. Without mandatory prior review, a company could launch a hiring system that discriminates or a medical assistant that recommends incorrect treatments. Altman asks for trust, but the data shows that the industry prioritizes the market over safety. Without controls, errors are discovered only after they have already caused harm.
Trust us, it's the seventh time we've promised 🔄
Altman asks us to believe him, just as we once trusted social networks, self-driving cars, and banking algorithms. It's always the same: first they ask for speed, then self-regulation, and when everything blows up, they ask for understanding and deadlines to patch it up. Altman's proposal sounds like freedom, but it smells like déjà vu. A billionaire asking for fewer controls is not innovation; it's the same old playbook.