Examine the legal implications and potential loopholes in new AI model licenses, particularly the OpenRAIL-M license used for models like BLOOM and Stable Diffusion. Explore the dilemma between open-source software principles and attempts to restrict harmful applications, discussing the challenges of defining "good" versus "bad" usage. Analyze a significant legal vulnerability in these licenses and introduce the OpenRAIL++ License as a potential solution. Clarify the distinction between these licenses and copyright law, and conclude with final thoughts on the complexities surrounding AI model licensing and responsible development.