Given the catastrophic risks associated with the misuse of AI technology, especially when it comes to nuclear weaponry, I strongly believe we should ban future open-source AI models that could be retrained for such purposes. The potential for harm far outweighs the benefits of open access in this context.