AntChain and Stanford University Unveil Open-Source Deepfake Datasets to Accelerate Global AI Security Advancements
2025-08-18 / Read about 0 minute
Author:小编   

At the International Joint Conference on Artificial Intelligence in Montreal, Canada, the workshop dedicated to deepfake detection technology garnered significant attention. During this prestigious event, AntChain unveiled the DDL-Datasets, comprising an impressive 1.8 million images. Simultaneously, Stanford University, in collaboration with other institutions, introduced the DeepAction dataset, which features 2,600 videos of human actions created by advanced models alongside their authentic counterparts. Furthermore, the conference introduced a groundbreaking multimodal semantic embedding-based AI-generated action detection technology, and the accompanying paper was honored with the Best Paper Award. Intriguingly, the research also revealed disparities in the ability to discern AI-generated fake videos among individuals from diverse cultural backgrounds. ZOLOZ, a prominent brand under the AntChain umbrella, showcased its substantial long-term investments and impressive global application achievements in the realm of AI security.