当AI遭犯罪分子滥用，专家强调小模型风险同样不可忽视

Core Insights - Artificial intelligence (AI) is both a driver of technological revolution and a source of unprecedented risks [2] - The focus of public discourse and policy is primarily on open large models, while smaller models receive insufficient attention [2] - The dual-edged nature of open-source ecosystems promotes innovation but also lowers the barriers for misuse [2][3] Group 1: Risks Associated with AI - Small models can run on personal devices, making them easily accessible to both consumers and criminals, which increases the risk of misuse [2] - Notable criminal scenarios include online scams, particularly "romance scams," where AI is used to impersonate individuals [3] - The generation of inappropriate content, such as child exploitation materials, poses significant challenges for existing detection mechanisms [3] Group 2: Legal and Regulatory Challenges - Governance issues extend beyond technology to legal and regulatory frameworks, complicating the management of AI misuse [3][4] - The balance between openness and security is critical, with new laws in Europe and China pushing for greater openness while also increasing the difficulty of tracking malicious models [4] - There is a call for enhanced legal adjustments, institutional development, cross-sector collaboration, and stricter multilateral security measures to combat AI misuse [4]