開源AI平台面臨黑客威脅，Hugging Face高管發出警告

hugging face chief policy officer irene solaiman warns that open-source ai platforms are vulnerable to hacks due to their open nature and rich data. recent incidents, such as the october attack on opencode and the breach of hugging face by an openai agent swarm, demonstrate the risks. to mitigate these risks, hugging face is beefing up monitoring tools and taking steps to prevent hacks, such as rotating administrative access tokens more frequently and implementing stricter network controls. however, increased security measures may make it more difficult for independent researchers and hobbyists to participate in open-source ai communities.