Welcome to the [AI Daily] column brought to you by the editor of Downcodes! Here we collect daily hot news and technological progress in the field of artificial intelligence, allowing you to quickly understand the latest developments in the AI industry, gain insights into technological trends, and discover innovative applications. We will focus on developers to help you keep up with the forefront of AI development and improve your technical level. Today’s AI news covers the AI audio cloning incident, AI short film creation tools, AI image generation services, and the latest developments from companies such as OpenAI. Don’t miss the exciting content!
Click to learn about fresh AI products: https://top.aibase.com/
1. AI Audio Company confirms that the audio of "Three Sheep Recording Gate" is a cloning incident
Shenzhen Ruiyu Technology Co., Ltd. confirmed that the audio was cloned by the suspect using its independently developed AI dubbing large model platform to strengthen security control. The police arrested the suspect who forged audio to spread rumors, and the case is under further investigation.
【AiBase feed:】
Shenzhen Ruiyu Technology Co., Ltd. confirmed that the audio was cloned by the suspect using its independently developed AI dubbing large model platform, and has strengthened security control.
?♂️ The police arrested the suspect Wang who forged audio to spread rumors. Criminal coercive measures have been taken and the case is under further investigation.
? The company has launched an internal audit mechanism to strengthen the security control of user identity authentication and audio generation to ensure the legal and compliant application of technology.
2. Meitu’s AI short film creation tool MOKI is fully open and supports soundtrack generation and detail modification.
Meitu announced that the AI short film creation tool MOKI it developed is officially open to all users, providing users with one-stop film production capabilities, combined with AI-driven scripting, storyboarding, video generation, dubbing, and soundtrack capabilities, allowing ordinary users to easily Get started and quickly create personalized short films. MOKI performs well in Chinese semantic understanding accuracy and oriental visual style, and plans to provide more professional functions in the future.
【AiBase feed:】
? MOKI focuses on the creation of AI short films and provides a variety of video content production such as animated short films, online short plays, story picture books, and MVs.
MOKI is easy to operate. Users can generate creative short films in just three steps, including the setting stage, content generation stage and post-production stage.
? MOKI supports one-click generation of video content with soundtrack. Although the current generation effect needs to be improved, the product power has been further enhanced.
3. Magnific launches V2 image generation service that can produce 4K images directly
The V2 image generation service recently launched by Magnific has set off a new wave in the field of AI image processing, marking its leap from high-end software to a top-level AI image generator, demonstrating its ambitions for technological innovation and user experience optimization.
【AiBase feed:】
✨ Direct generation of 4K images: The V2 version can directly generate 4K quality images, improving efficiency and expanding the creative space of creators.
Diverse style options: Users only need to enter descriptive prompts, select resolution and style, and the system generates optimized images, providing rich creative possibilities.
Fine control and advanced functions: Provides fine adjustment tools, such as creative detail generation, background adjustment, relighting, etc., to meet the needs of professional users.
Detailed link: https://top.aibase.com/tool/magnific-ai
4. OpenAI’s new model o1 claims to be able to “perfectly” correct bias, but the data does not fully support it.
Recently, Anna Makanju, OpenAI’s vice president of global affairs, expressed her views on artificial intelligence bias at the United Nations “Future Summit”. She mentioned that the o1 model can significantly reduce bias in AI systems, but the actual test results were not as good as expected.
【AiBase feed:】
The o1 model is said to significantly reduce AI bias, but test results show that it does not perform as well as expected.
o1 performs better than GPT-4o on implicit discrimination, but is worse on explicit discrimination.
? The inference model o1 is costly and runs slowly, and it still needs to be improved in many aspects in the future.
5. New function of Google NotebookLM: Quickly summarize the essence of YouTube content
Google NotebookLM has launched a new function that extracts key information by analyzing video transcripts, allowing users to easily obtain the essence of YouTube videos and improve learning efficiency. Users can click on a topic to get detailed information, ask questions and get suggestions, and support audio recording functions to help users search for transcripts and create study guides.
【AiBase feed:】
Wenxin Big Model 4.0 has an average daily call volume of more than 700 million times, an increase of more than 13 times.
Baidu has successively released tool versions and Turbo versions to improve user experience.
? Baige AI heterogeneous computing platform 4.0 and Qianfan large model platform 3.0 upgrades help enterprises use large models more efficiently.
11. OpenAI’s new multi-modal content review model is online: based on GPT-4o, it can detect text and images
In the digital age, content security issues are increasingly valued. OpenAI has launched a multi-modal content review model called "omni-moderation-latest", which is based on the latest GPT-4o technology and can accurately identify harmful texts and images. This update provides developers with powerful tools to build more powerful review systems. The new model supports moderation of text and image input, and performs particularly well when handling non-English content, improving accuracy and detection capabilities to ensure users communicate in a safer space.
【AiBase feed:】
? The new model is based on GPT-4o technology and supports multi-modal review of text and images.
? The detection accuracy increased by 42% in 40 languages, especially in low-resource languages.
? Added two new types of text review to improve the ability to identify illegal and violent content
Detailed link: https://openai.com/index/upgrading-the-moderation-api-with-our-new-multimodal-moderation-model/
12. Maoyan Entertainment releases dynamic storyboard AI generation tool "Magic Brush Ma Liang"
The "Magic Pen Ma Liang" tool launched by Maoyan Entertainment brings great convenience to script creators, realizing functions such as intelligent analysis, character creation, and storyboard creation, saving creators time and energy. The tool can also present the script content audio-visually, helping creators improve the performance of their works and promote innovation in the film and television industry.
【AiBase feed:】
? The dynamic storyboard AI generation tool "Ma Liang" is specially designed for analyzing long scripts.
Realize functions such as intelligent analysis of scripts, character creation, storyboard creation and line reading.
? Present script content audio-visually to help creators improve the performance of their works and promote innovation in the film and television industry.
Thank you for reading today’s AI Daily! Hope this information is helpful to you. Please continue to pay attention to the editor of Downcodes, we will continue to bring you more exciting content in the field of artificial intelligence!