OpenAI strengthens GPT-5.6 to protect against a hint of an attack

OpenAI announced that it would strengthen its GPS-5.6 model through the internal AI Red Corps in order to resist the warning injection. This enhancement could significantly enhance AI security in financial applications and reduce the risk of unauthorized operations。