Xinhua deepfake news anchors stir scepticism, distrust
A pair of virtual Chinese news readers raised concerns about their use by the Chinese authorities for propaganda and disinformation. Developed by Chinese news service Xinhua in collaboration with search engine company Sogou, " Qiu Hao" a nd " Zhang Zhao" were created using facial recognition, deep learning, 3D reconstructions, machine translation and facial expression modeling to mimic the appearance and voice of real human anchors. Xinhua claimed the anchors, one of which s peaks English and the other Mandarin , “can read texts as naturally as a professional news anchor”. Both were put to work on Xinhua's WeChat channel and were said to be intended to reduce news production costs and improve efficiency.” A Sogou manager told SixthTone that the new technology would "liberate the labour force". However, the anchors met with skepticism from the public, with many questioning whether they seem as real as human reporters. Meantime, others commented that the anchors' voices sound stiff and have trouble with pauses. Experts warned that while the technology is impressive, it is essentially just a digital puppet that reads a script as opposed to an intelligent system. Concerns were also epxressed about the potential misuse of the technology to spread Chinese government propaganda and misinformation. Deepfake Deepfakes (a portmanteau of ' deep learning ' and 'fake') are images, videos, or audio which are edited or generated using artificial intelligence tools, and which may depict real or non-existent people. They are a type of synthetic media. Source: Wikipedia 🔗 System 🤖 Qiu Hao Zhang Zhao Operator: Xinhua users Developer: Sogou Country: China Sector: Media/entertainment/sports/arts; Politics Purpose: Read news Technology: Deepfake - video, audio; Machine learning Issue: Ethics/values; Mis /disinformation; Robustness
- Date it happened
- 2018-11-01
- Organisation involved
- Xinhua
- Product, system or model
- Qiu Hao; Zhang Zhao
This incident was imported from AIAAIC and is used under CC BY-SA 4.0. Our additions to it — the structured fields, the translation, the checks against other reports — are published under the same licence.
This is a record of what was reported, not a finding that anyone broke the law. If it names your organisation and you believe it is wrong, the corrections process is free and open to everyone.