OpenAI Prioritizes Speed Over Safety in AI Race

OpenAI is accelerating the release of new AI models, significantly reducing safety testing time, according to the Financial Times. While GPT-4 underwent six months of risk evaluation, its successor had only a few days. Tests are now more superficial and often done on earlier versions, not final models. Especially concerning are gaps in misuse testing, such as for bio-weapon risks. The upcoming EU AI Act could soon restrict such practices.