Welcome to the 【AI Daily】 column! Here is your daily guide to exploring the world of artificial intelligence. Every day, we present you with hot topics in the AI field, focusing on developers to help you gain insights into technological trends and understand innovative AI product applications.
Fresh AI products Click to learn more: https://top.aibase.com/
1. The National Radio and Television Administration issues a "Management Notice" to address AI "magic modifications," with classic films being parodied.
The National Radio and Television Administration has issued a management notice regarding the phenomenon of AI "magic modifications" in videos, emphasizing governance measures to protect classic culture. Recently, several classic films and TV dramas have been parodied, raising public concern about traditional culture. The management requires provincial bureaus to urge short video platforms to clean up related content and implement content reviews for generative AI to prevent misleading information and copyright infringement.
【AiBase Highlights:】
🚫 The National Radio and Television Administration has issued a notice requiring the management of AI "magic modifications" in videos to protect classic culture.
🎥 Several classic films and TV dramas have been parodied, affecting public perception of traditional culture.
🔍 Platforms need to strictly review generative AI content to avoid misleading and copyright infringement.
2. Hailuo AI launches an overseas version with a powerful AI voice cloning feature.
Hailuo AI recently launched its overseas version of the audio cloning module, allowing users to perfectly replicate their voices with just 10 to 60 seconds of audio samples. This technological breakthrough has garnered widespread attention in the field of Chinese voice cloning, with many users expressing surprise at its excellent audio cloning results. The system is not only easy to use but also supports multiple languages and emotional selections, greatly satisfying user needs.
【AiBase Highlights:】
🌟 With just 10 to 60 seconds of audio, Hailuo AI can perfectly replicate user voices, breaking through the bottleneck in Chinese voice cloning.
🎤 The system supports 12 languages, allowing users to choose different emotions for richer voice expressions.
💻 Currently, Hailuo AI's audio features are free to use, allowing users to easily create multiple voice models.
Details link: https://www.hailuo.ai/audio
3. Runway Act One update: Seamlessly integrate your performance and voice into video characters.
The latest update of Runway Act One brings revolutionary changes to video production, allowing users to directly apply their performances and voices to other video characters for perfect synchronization. This technological breakthrough not only lowers the creative threshold but also provides unprecedented flexibility for film and television creation, enabling creators to audition and transfer performances anytime, anywhere.
【AiBase Highlights:】
🎭 Multi-dimensional performance transfer: Actions, voices, and expressions can be seamlessly transplanted onto any character.
📱 Convenient auditions: Actors can easily film with their phones and transfer performances to target characters.
🖥️ AI integration: Using tools like Midjourney, creators can extend short videos into complete performances.
Details link: https://top.aibase.com/tool/runway
4. OpenAI is about to launch the new Sora video generator, supporting multiple generation methods.
OpenAI recently announced at the C21Media conference in London the upcoming release of the updated Sora video generator, which will support various generation methods using text, images, and videos, greatly enhancing the user experience in video creation. The new version has significant improvements in efficiency and speed, expected to be officially released during the winter promotion in December, along with potential new features like GPT-4.5.
【AiBase Highlights:】
🌟 The updated Sora video generator will support text, image, and video generation, enhancing creative flexibility.
🚀 The new generator has significant improvements in speed and efficiency, enhancing user experience.
📅 Expected to be released during the winter promotion in December, possibly alongside new features like GPT-4.5.
5. The ultra-high-definition video restoration tool VISION XL makes blurry videos clear with one click.
With the advancement of technology, VISION XL stands out as a video restoration and super-resolution tool with its excellent performance and ease of use. It can not only repair missing parts of videos and remove blurriness but also significantly enhance video clarity, achieving up to four times super-resolution. Its processing framework based on latent diffusion models reduces dependence on additional pre-training modules, greatly improving the efficiency of high-resolution video processing.
【AiBase Highlights:】
✨ VISION XL can repair missing parts of videos, remove blurriness, and enhance clarity, achieving up to four times super-resolution.
⚙️ Using a processing framework based on latent diffusion models reduces reliance on additional pre-training modules, improving processing efficiency.
🚀 Only requires 13GB of video memory to process 25 frames of video, with a processing time of no more than 2.5 minutes, suitable for rapid application scenarios.
Details link: https://vision-xl.github.io/
6. Elon Musk's social media platform X launches the image generator Aurora.
Elon Musk's social media platform X has recently launched a new image generator called Aurora, aimed at creating photorealistic images. Although some users were unable to access this feature shortly after its launch, Aurora still allows users to generate images of public and copyrighted characters, including Mickey Mouse, without restrictions. The tool performs excellently in generating still life and landscape images but also has some shortcomings, such as unnatural merging of objects and missing fingers in portraits.
【AiBase Highlights:】
🌟 The new image generator Aurora is live, allowing users to generate various images.
🚫 Some users were unable to access this feature within hours of its launch.
💰 The X social platform has opened the Grok feature to all users.
7. Google's new Gemini-Exp-1206 model sweeps competitors, surpassing ChatGPT to become the new AI king.
Google's latest Gemini-Exp-1206 model has garnered widespread attention in the generative AI field, achieving a high score of 1379 on the LMArena leaderboard, surpassing ChatGPT-4.0's score of 1366, showcasing its outstanding overall capabilities. Although Gemini-Exp-1206 performs exceptionally well in various assessments, it still lags behind ChatGPT-4.0 in the number of votes, indicating the latter's advantage in reliability.
【AiBase Highlights:】
🌟 Gemini-Exp-1206 scored 1379 on the LMArena leaderboard, surpassing ChatGPT-4.0's score of 1366.
🗳️ ChatGPT-4.0 received 21,929 votes, significantly higher than Gemini-Exp-1206's 5052 votes, demonstrating its reliability.
🔍 The Gemini experimental model offers developers an unprecedented AI experience opportunity but is still in the testing phase and not suitable for production use.
Details link: https://ai.google.dev/gemini-api/docs/models/experimental-models?hl=en
8. NegToMe redefines image generation: Reducing copyright risks, enhancing diversity, and improving visual effects.
NegToMe is a groundbreaking image generation technology that utilizes image-driven adversarial guidance methods, breaking through the limitations of traditional negative prompts and significantly enhancing the diversity and quality of generated images. It addresses copyright protection issues by reducing the similarity of generated content to copyrighted works while also performing excellently in cross-domain applications, providing creators with greater creative freedom.
【AiBase Highlights:】
🎨 NegToMe significantly enhances the diversity of generated images through image-driven adversarial guidance methods, particularly excelling in handling race and gender.
🔒 This technology reduces the similarity of generated content to copyrighted works, with tests showing a 34.57% reduction in similarity, effectively addressing copyright protection issues.
⚙️ NegToMe is easy to integrate, requiring only minimal code from developers, with inference time remaining virtually unaffected, compatible with various diffusion models.
Details link: https://github.com/1jsingh/negtome
9. X opens Grok AI to all users, allowing regular users to access generated images for free.
xAI recently announced that its chatbot Grok is now available to users worldwide, providing a low-cost AI experience opportunity. Users in the free version face some usage restrictions, such as limits on the number of images generated and messages sent daily. This move not only attracts more users to learn about AI technology but also reflects xAI's commercial strategy in promoting its products.
【AiBase Highlights:】
🖼️ Grok allows users to create or analyze up to 3 images per day.
💬 Users can only send 10 messages within two hours to control usage frequency.
📈 xAI attracts users by making Grok free, with potential for more paid features in the future.
10. Google Photos launches the 2024 annual photo review: AI intelligently generates and records your memorable moments.
With the development of digital technology, Google Photos has launched the 2024 annual photo review feature, using AI technology to provide users with a personalized experience. Through Gemini AI, users can receive intelligently generated photo captions, reviewing important moments and shooting data. Although this feature offers users the opportunity to share beautiful memories, it may also evoke some unpleasant memories.
【AiBase Highlights:】
🤖 AI technology generates personalized photo captions, highlighting important moments of the year.
📊 Provides detailed shooting data statistics, making it easy for users to share personal metrics.
😢 May evoke some unpleasant memories, as AI has not yet fully understood users' emotional needs.
11. OpenAI decides to collaborate with military contractors, facing opposition from internal employees!
OpenAI's collaboration with Anduril has sparked strong reactions from employees, many of whom are concerned about the use of technology in military applications, demanding more transparency. Despite management emphasizing that the collaboration is limited to defense systems, employees are skeptical about this boundary.
【AiBase Highlights:】
🌐 OpenAI's collaboration with Anduril raises employee concerns about AI military applications.
🛡️ Management emphasizes that the collaboration is limited to defense systems, but employees express skepticism about the limitations of technology applications.
📉 Policy shifts indicate that OpenAI is beginning to accept the application of its technology in military fields.
12. AI experts: One ChatGPT query is equivalent to wasting half a liter of water.
The rapid development of generative artificial intelligence has brought environmental issues, particularly regarding energy and water resource consumption. Professor Kate Crawford pointed out in a lecture that without sustainable measures, the energy consumption of generative artificial intelligence could reach levels comparable to Japan within a year.
【AiBase Highlights:】
🌍 One ChatGPT query wastes half a liter of water, reminding people to pay attention to the impact of artificial intelligence on water resources.
⚡ The energy consumption of generative artificial intelligence could reach levels comparable to Japan within a year, necessitating the development of sustainable plans.
🤝 Sustainability should be the top priority in the AI industry, rather than competition rankings.
