Chinese tech giant Tencent Cloud has developed a platform for generating realistic bots, referred to as “digital people.” The service uses a proprietary artificial intelligence (AI) system to create videos of any individual based on a three-minute moving image and 100 sentences of speech material. Users can choose from five digital human variants: 3D realistic, 3D semi-realistic, 3D cartoon, 2D real person, and 2D cartoon. Background and hue can be customized, and the videos created reportedly avoid the flat intonation and uniform speech rhythm often associated with conventional acoustic models.
The digital figures are marketed for use in live-streamed commercials, e-commerce, and other applications. However, the risk of misuse is high, with the realistic-looking bots being increasingly used for deepfakes. Multimodal AI models in the form of deepfakes present a potential risk of generating images, videos, or voices that deceive viewers into thinking they are real. Twitter CEO Elon Musk and Europol have both warned of the dangers of deepfakes being used for criminal activity.
Chinese regulations are in place to prevent such scenarios, but the development of such bot technology highlights the increasing use of AI in creating misleading content. Tencent aims to build an AI-driven “intelligent digital human factory,” relying on a self-service platform for production, sales, and service of the bots.