Meta Corporation is planning to release a new version of artificial intelligence called "Llama 3" in July. The design of this new AI is intended to provide more flexible answers to various questions, including controversial ones such as "how to win a war." It is reported that the developers at Meta hope the new version of Llama can address some of the issues that Llama 2 currently cannot handle. However, this plan has raised some concerns. Similar to Google's Gemini, Meta is also attempting to train its AI model to avoid answering potentially dangerous questions. This is to prevent the AI from occasionally providing dangerous information, such as how to win a war or create a bomb, when experiencing hallucinations. Previously, Google's Gemini faced controversy for generating diverse but historically inaccurate images. Therefore, Meta's new move undoubtedly surprises people. After all, both Google and Meta are striving to ensure that their AI models do not answer potential dangerous questions. However, Meta's plan has also raised some doubts. Some people are concerned that if the AI model is too flexible, it may lead to unforeseen consequences. For example, if the AI model is used to answer controversial questions, it may cause social unrest or mislead users. Overall, Meta's new plan undoubtedly brings new challenges and opportunities to the AI industry. On the one hand, this may drive the development of AI technology, enabling it to better serve humanity. On the other hand, it also reminds people to be more cautious in dealing with AI technology, ensuring that it does not bring potential dangers while advancing.