AI chatbots, powered by large language models, face challenges in alignment and safety. Researchers and regulators work to mitigate risks through enhanced training, systematic defenses, and oversight.