Core Viewpoint - The rise of AI-generated "New Year greetings" featuring celebrities has sparked discussions about the legal boundaries and potential infringements on rights associated with unauthorized use of likenesses and voices [1][5]. Group 1: AI Technology in New Year Greetings - AI technology is being utilized to create engaging short videos featuring familiar celebrities, athletes, and historical figures, which has become popular on social media platforms as the Spring Festival approaches [1][2]. - The core technology behind these videos includes deep synthesis techniques, specifically visual synthesis (deep fake) and voice synthesis, which together create realistic AI-generated content [3][4]. Group 2: Legal and Compliance Concerns - There are significant concerns regarding the compliance and safety of deep synthesis videos, particularly regarding the unauthorized use of individuals' likenesses and voices, which can infringe on personal rights [5][6]. - The Chinese Civil Code explicitly protects portrait and voice rights, and unauthorized deep synthesis can violate these rights, potentially leading to defamation or copyright infringement [6][7]. Group 3: Industry Response and Regulation - Some AI service platforms are implementing measures to prevent misuse, such as mandatory AI-generated content labels and user agreements prohibiting unauthorized use of likenesses and voices [7]. - There is a call for network platforms to take responsibility for managing the production and dissemination of such content, including establishing comprehensive management mechanisms and responding promptly to infringement notifications [7].

Core Viewpoint - Tongyi Qianwen announced an upgrade to its voice synthesis model Qwen3-TTS, which now supports multiple voice tones, languages, and dialects, accessible via Qwen API [1] Group 1: Product Features - Qwen3-TTS offers over 49 high-quality voice tones, catering to various gender, age, regional characteristics, and character settings to meet diverse scenario needs [1] - The model supports 10 major languages including Chinese, English, German, Italian, Portuguese, Spanish, Japanese, Korean, French, and Russian, with an average word error rate (WER) better than MiniMax, ElevenLabs, and GPT-4o-Audio-Preview [1] - Additionally, it supports dialect generation for languages such as Minnan, Cantonese, and Sichuan dialect [1]

Core Viewpoint - Alibaba's Tongyi Qwen has upgraded its flagship voice synthesis model Qwen3-TTS, enhancing its capabilities in multi-tone, multi-language, and dialect support [1] Group 1 - Qwen3-TTS now features richer voice options and improved support for multiple languages and dialects [1] - The model's prosody and speech rate have been made more natural and human-like [1] - Qwen API access is available for users to utilize the upgraded model [1]

Core Insights - Alibaba's Qwen3-TTS has undergone a comprehensive upgrade, enhancing its capabilities in voice synthesis [1] Group 1: Product Features - Qwen3-TTS is a flagship voice synthesis model that supports multiple voice tones, languages, and dialects [1] - Key improvements include richer voice tone support and continuous enhancement of multilingual and dialect capabilities [1] - The model now features more natural prosody and speech speed, making it more human-like [1]

Group 1 - The Ministry of Education announced that as of June 20, 2025, there are a total of 3,167 higher education institutions in China, including 2,919 regular higher education institutions, 1,365 undergraduate schools, and 1,554 vocational colleges [1] - The Ministry of Education warned candidates and parents to verify information through official channels during the upcoming college admissions process, as there have been instances of false information and anxiety-inducing tactics used by some individuals and organizations [2] - Beijing has launched a three-year initiative to enhance the quality of ideological and political education teachers in schools, aiming to build a well-structured and high-quality teaching team [3] Group 2 - BOSS Zhipin issued a proposal to address the issue of false job postings in the recruitment market, urging industry peers to take responsibility and improve service standards while promoting transparency and establishing a credit evaluation system [4] - iFLYTEK upgraded its voice synthesis technology, achieving breakthroughs in voice replication and hyper-realistic synthesis, supported by a three-stage hierarchical voice modeling framework [5]