AI Toolset adopts a monthly 3-issue layout collection, updating the latest AI industry news daily. This issue is the collection of AI industry news for June Issue 2 (June 11 to June 20).
June 20
- The Cyberspace Administration of China released information on algorithm filing for domestic deep synthesis services. Article 19 of the Provisions on the Administration of Deep Synthesis of Internet Information Services clearly stipulates that providers of deep synthesis services with public opinion attributes or social mobilization capabilities shall, in accordance with the Provisions on the Administration of Algorithm Recommendations for Internet Information Services, fulfill the filing, change, and cancellation of filing procedures. (Read full article:ITHome）
- At today's Changhong AI Large Model TV launch event, Changhong TV's AI large model "Changhong Super Brain" was officially released. According to the introduction, Changhong Super Brain has three core capabilities: perception, understanding, and creation. It can conduct natural and continuous conversations with users and can also create content and services based on users' personalized needs. (Read full article:Changhong TV Official Account）
- According to Tech Planet, recently, Kuaishou's video editing product "Kuaiying" and feature shooting product "Yitian Camera" have recently been beta testing multiple AIGC creation features, including "AI Anime Video," "AI Copywriting Recommendations," "AI Instant Universe," and "AI Stick Figure." (Read full article:Tech Planet Baijiahao）
- SPACE10, the research and design lab under Swedish furniture giant IKEA, recently launched an AI-designed next-generation sofa, aiming to introduce eco-friendly and portable concepts and change people's perception of traditional sofas. (Read full article:PingWest）
- According to a report published earlier this month by market research firm Resume Builder, in 2024, 43% of corporate recruitment will introduce artificial intelligence to simplify the recruitment process and improve interview efficiency. (Read full article:ITHome）
- Autonomous driving company TuSimple successfully completed China's first fully driverless test of an autonomous heavy truck on public roads, with no safety driver on duty and no human intervention throughout. The fully driverless test area includes public autonomous driving test roads such as the Deepwater Port Logistics Park and the East China Sea Bridge, with a total route of about 62 kilometers. (Read full article:QbitAI）
- Liangsheng Digital, a cinematic digital special effects design company, released AINIA, the first AI intelligent life form created based on NVIDIA Omniverse. Liangsheng Digital positions AINIA as a superstar singer. It has mastered the singing techniques of more than 100 historically famous singers and has a library of tens of thousands of songs. In the future, it will also be able to write lyrics and compose music on its own. (Read full article:36Kr）
June 19
- At yesterday's Meitu Imaging Festival, Meitu released its own "AI full suite": the visual large model MiracleVision, as well as multiple AI applications—image generation (WHEE), oral broadcast video production (Kaipai), video editing (WinkStudio), commercial design (Meitu Design Studio 2.0, overseas version X-Design), and digital human generation (DreamAvatar). (Read full article:36Kr）
- At today's Tencent Cloud Industry Large Model and Intelligent Application Technology Summit, Tencent Cloud disclosed the progress of its industry large model solutions and announced the MaaS (Model-as-a-Service) capability panorama. Tencent Meeting is about to launch an AI assistant covering the entire meeting process. (Read full article:Tencent）
- Microsoft Asia Research (MSRA) responded that Microsoft will establish a new laboratory in Vancouver, Canada, confirming that it will include Microsoft's AI research experts in China in the new laboratory. At the same time, MSRA denied that it will withdraw from China. (Read the full article:TMTPost）
- The UK government announced that tech entrepreneur Ian Hogarth will lead a new AI task force to study the security risks brought by artificial intelligence. In April of this year, the UK government pledged to invest 100 million pounds to establish the task force. The task force will study the risks surrounding artificial intelligence and conduct safety research. (Read the full article:Sina Finance）
- At the world's largest computer vision conference CVPR 2023 held yesterday in Vancouver, Canada, LG AI Research unveiled the company's first generative AI commercialization service, Captioning AI, providing a more efficient solution for customers who need to manage large numbers of images. (Read the full article:ITHome）
- General large model startup Beijing Shengshu Technology Co., Ltd. announced the completion of nearly 100 million yuan in angel round financing. This round of financing was led by Ant Group, with BV Baidu Ventures and Zhuoyuan Capital participating, and the current valuation has reached 100 million US dollars. (Read the full article:National Business Daily）
- AI + cloud industrial software company '3vjia' completed hundreds of millions of yuan in Series D financing, with investors including Guoke Investment and Huichuan Technology, and the post-investment valuation will reach nearly 600 million US dollars. (Read the full article:36Kr）
June 18
- Researchers at Princeton University introduced in a paper a realistic 3D scene procedural generator for the natural world, 'Infinigen'. Infinigen is completely procedural; from shapes to textures, everything is generated from scratch through random mathematical rules, and it can easily customize, simulate, and generate realistic situations for various specific tasks. (Read the full article:NetEase News）
- At yesterday's Li Auto Family Tech Day press conference, Li Auto said it will adopt large model algorithms, develop its own cognitive large model Mind GPT, and introduce AI technology into the in-car assistant 'Li Xiang Tong Xue'. (Read the full article:Tencent News）
- Xingzheng Global announced recently that its AI trader has officially gone live, making it the first fund company to apply AI technology to the field of capital trading. Industrial Fund also officially announced the self-developed 'Xing Xiao Er' AI bond trading robot, making Industrial Fund the first public fund company to launch an intelligent inquiry robot on the iDeal platform of the China Foreign Exchange Trade System. (Read the full article:The Paper）
- 'Surreality: AI Dali + Dream Dali' immersive art exhibition opened yesterday at the Shanghai Himalayas Museum. This major exhibition integrates the artist Dali's way of thinking with AI technology, reproducing the symbols and expressive techniques in Dali's art through artificial intelligence technology. (Read the full article:Shanghai Securities News China Securities Network）
- Gannett, the largest newspaper group in the United States, plans to introduce generative artificial intelligence into its news publishing system to save costs and improve efficiency. But the group said it will involve human editors to prevent artificial intelligence from automatically publishing without supervision. (Read the full article:ITHome）
June 17
- The Recording Academy of the United States announced that it is revising the Grammy Award rules, clearly stating that all Grammy Award categories will not accept works created entirely by non-human creators; only works by 'human creators' are eligible to win. (Read the full article:ITHome）
- Meta launches Voicebox, a large speech generation model product, capable of synthesizing speech in six different languages including English, French, German, and Spanish. In addition to generating speech content from text, it can also perform tasks such as denoising, content modification, and style transfer. (Read the full article:The Paper）
- While attending the Viva Technology exhibition in Paris, France, Musk said that artificial intelligence, as "the most disruptive technology in history," may produce positive results, but "there are definitely risks to the public," and this technology should be regulated. (Read the full article:Shangguan News）
June 16
- The Institute of Automation of the Chinese Academy of Sciences released in Shanghai this morning the new-generation artificial intelligence large model developed by the institute - Zidong Taichu 2.0. Compared with the first generation, the new large model has focused on improving decision-making and judgment capabilities, achieving a leap from perception and cognition to decision-making, and will play a greater role in the future in fields such as healthcare, transportation, and industrial production. (Read the full article:The Paper）
- WeChat Channels released recommendations on creators' standardized use of deep synthesis technology to generate content, requiring conspicuous labeling of non-real audio and video content generated or synthesized using virtual reality, deep synthesis technology, and the like. (Read the full article:WeChat Video Creation Safety Center）
- Meta's Chief AI Scientist and Turing Award winner -Yann LeCunsaid yesterday at the Viva Tech conference held in Paris that AI's current intelligence level is not even as good as a dog's, and it should not be regarded as a threat to humanity. (Read the full article:IT Home）
- The "Artificial Intelligence Framework Ecosystem Summit 2023" was held in Shanghai. During the summit, the "Joint Initiative on Building the Artificial Intelligence Framework Ecosystem" was released, the Shanghai MindSpore AI Framework & Large Model Innovation Center was launched with the first batch of partners moving in, and the MindSpore Community Council was established. (Read the full article:Geek Net）
- Mercedes-Benz and Microsoft announced on Thursday that the two companies are collaborating to test in-car ChatGPT, which can be used in more than 900,000 vehicles in the United States. The technology will be used through the in-car voice command "Hey Mercedes," and is expected to greatly expand the system's functions. (Read the full article:Zhitong Finance）
- Kunlun Wanwei announced yesterday that its controlled subsidiary Star Group plans to issue shares to acquire all equity of Singularity AI, with a transaction consideration of about 1.1 billion yuan. Behind Singularity AI is Singularity Zhiyuan, the developer of China's ChatGPT-like "Tiangong". (Read the full article:36Kr）
- The United Nations Industrial Development Organization announced that in July it will establish a Global Alliance on Artificial Intelligence for Industry and Manufacturing with Huawei and other international enterprises, committed to building a platform for mutual collaboration and knowledge sharing, promoting innovation and application of artificial intelligence technology in the industrial and manufacturing fields, and advancing the development of the digital economy. (Read the full article:Huawei official website）
June 15
- Xiaoice and NetEase Cloud Music announced a strategic cooperation. The AI singer music creation software jointly launched by both parties, "NetEase Cloud Music · X Studio", has officially launched and is available for free to the vast number of musicians and music lovers. (Read the full article:AI Xiaoice Official Account）
- "Baichuan Intelligence," founded by former Sogou CEO Wang Xiaochuan, today officially launched a 7-billion-parameter Chinese-English pretrained large model - baichuan-7B. It has been released on platforms such as Hugging Face, GitHub, and ModelScope, and has achieved the best results on multiple evaluation leaderboards. (Read the full article:Xinzhiyuan）
- iFlytek announced that the iOS beta version of the Spark Cognitive Model has officially launched on the Apple iOS platform. Currently, using this App requires "beta qualification", and you need to submit an application on the iFlytek official website. Only after approval can you participate in the beta and log in to use it. (Read the full article:iFlytek Official Account）
- The online code hosting platform GitHub recently conducted a survey of U.S. enterprises with more than 1,000 employees. The questionnaire surveyed 500 enterprise developers, and the results showed that 92% of programmers are using various AI tools, and 70% of programmers believe that the introduction of AI tools has significantly improved work efficiency. (Read the full article:ITHome）
- Sinopec and Baidu signed a strategic cooperation framework agreement. The specific cooperation content includes industry artificial intelligence infrastructure construction, Internet+ new business formats, digital transformation and upgrading, and innovation in new technologies such as large models. The two parties will also cooperate in service sectors such as supply chain, automobiles, insurance, and lifestyle, and explore the use of technologies such as AIGC and digital humans to help build Sinopec's smart media platform. (Read the full article:Guangming Online）
- Google's parent company Alphabet has advised employees not to input confidential materials into AI chatbots, and the company has also reminded IT engineers to avoid directly using computer code generated by chatbots. (Read the full article:ITHome）
- A report by the consulting firm McKinsey titled "The Economic Potential of Generative AI" shows that by around 2045, 50% of jobs will be taken over by AI. (Read the full article:ITHome）
- Changhong TV announced today that some of its products will be equipped with the "Changhong Super Brain" artificial intelligence system and will officially enter public beta for users. On the 20th of this month, Changhong will hold the industry's first artificial intelligence large-model TV launch event online, introducing the "Changhong Super Brain" product and conducting a demonstration experience. (Read the full article:Changhong TV Official Account）
June 14
- OpenAI today announced a major update to its large language model API (including GPT-4 and gpt-3.5-turbo), including adding function calling capabilities, reducing usage costs, and providing a 16,000-token version for the gpt-3.5-turbo model. (Read the full article:QbitAI）
- Recently, Microsoft launched an introductory course series for AI beginners, "AI for Beginners", and it is completely free. It includes 24 lessons in total. In addition to drawing "knowledge point comics" for each topic, it also includes very rich and diverse quizzes and practical projects to ensure that learners have the opportunity to practice. (Course address:AI for Beginners）
- The well-known no-code website design and publishing platform Framer has newly launched Framer AI feature. Users only need to input a text description, and it can automatically design, lay out, generate, and publish a high-appearance responsive website page. (Read the full article:Framer Updates: Start with AI）
- The online document collaboration platform Coda (one of Notion's competitors) has officially launched Coda AI, similar to Notion AI. This AI assistant can help users draft article content, generate tables, find information, apply data, summarize meeting minutes, and more. (Website address:Coda AI）
- According to the latest developer survey report released by the programming Q&A website Stack Overflow, although programmers are skeptical about the accuracy of various AI-driven tools, this does not prevent them from using tools such as ChatGPT during development. 70% of respondents said they have already used AI tools for programming this year. (Read the full article:ITHome）
- Paul McCartney, a member of The Beatles, said he is using artificial intelligence to create the final Beatles song. This unnamed song is expected to be released later this year, with John Lennon's vocals aided by AI. (News source:NBC News）
- According to The Wall Street Journal, Microsoft did not heed OpenAI's warnings before using an unreleased GPT-4 model to power Bing's chat feature, leading to inaccurate and unpredictable responses from the chatbot. (Read the full article:IT之家）
- AI digital human video generation platform Synthesia raised approximately $90 million in its latest funding round, led by Accel, with Nvidia also joining the financing, bringing its valuation to $1 billion and officially promoting it to a "unicorn" company. (Read the full article:Cailian Press）
- Meta has recently open-sourced an AI model called I-JEPA, which can use background knowledge about the real world to fill in missing image segments, analyzing and completing incomplete images more accurately than existing image models. (News source:Reuters）
- A team of Microsoft researchers recently introduced Orca, a 13-billion-parameter model that can learn complex explanation traces and step-by-step reasoning processes from GPT-4. This method significantly improves the performance of existing state-of-the-art instruction-tuned models and addresses challenges related to task diversity, query complexity, and data scaling. (Read the full article:MarkTechPost）
June 13
- Microsoft's new Bing beats ChatGPT to support GPT-4 multimodal image recognition. It can not only understand memes and write web pages, but also "diagnose illnesses." This feature is currently a test feature available only to a small number of users and may be gradually rolled out to everyone later. (Read the full article:ifanr）
- As one of the most authoritative scientific journals, Nature recently made a clear statement: it prohibits the use of images and video content created by generative artificial intelligence (AIGC)! This means that, except for articles whose subject is discussing AI, any work accepted by Nature must ensure that there is no AIGC-generated or enhanced visual content. (Read the full article:QbitAI）
- iReader Technology announced that its first AI product, "Yue'ai Liao," has entered closed beta testing. This is the first conversational AI application in China's reading industry, supporting both role and reading modes. Users can either find a character from a book to converse with, or find a book to read through a character. (Read the full article:IT之家）
- 360 officially released the new AI product 360Zhinaolarge model. This large model is 360's self-developed hundred-billion-parameter large model. It already has cross-modal generation capabilities and possesses four major capabilities: text, image, speech, and video processing. It can be applied to scenarios such as text-to-image, image-to-image, and text-to-video. (Read the full article:ZAKER News）
- Salesforce launched the AI Cloud suite, bringing generative AI into multiple products under its umbrella, including its collaborative office application Slack, data analysis software Tableau, and automation process tool Flow, providing a total of 9 models to meet customers' various office and sales needs. (News source:BusinessWire）
- In a speech at London Tech Week, UK Prime Minister Rishi Sunak revealed that Google DeepMind, OpenAI, and Anthropic have committed to providing the UK government with "early or priority access" to AI models for research and safety purposes. (News source:Engadget）
- According to a report by Japan's Sankei Shimbun, Tokyo Governor Yuriko Koike announced today that all government departments in Tokyo will introduce generative AI represented by ChatGPT, and the plan will be fully implemented starting in August. (Read the full article:IT Home）
- According to research published by the American Psychological Association, employees who frequently interact with artificial intelligence systems are more likely to feel lonely, which may lead to insomnia and increased alcohol consumption after work. Researchers also found that employees who frequently use artificial intelligence systems are more likely to help colleagues, but this response may be triggered by their loneliness and social needs. (News source:American Psychological Association）
June 12
- Visual China's vcg.com announced the launch of an AI inspiration drawing feature, namely "generating images based on prompts." It is said that users can generate "high-quality, exquisite image content" through the AI inspiration drawing feature, thereby gaining more inspiration and creativity. (Read the full article:IT Home）
- At the 25th Shanghai International Film Festival "Sci-Fi Film Week," Guo Fan, director of The Wandering Earth, said he had been "anxious for a long time" about artificial intelligence represented by ChatGPT, but also revealed that his team is currently categorizing multiple artificial intelligence applications in the hope of applying AI in creation. (Read the full article:China Business Journal）
- OpenAI CEO Sam Altman met with SK Networks President Choi Sung-hwan in Seoul to discuss investment cooperation in artificial intelligence companies. OpenAI is evaluating the matter of opening a Korean office. (Report source:Jiemian News）
- Bloomberg reported that French President Macron met privately on the 9th with French Minister for Digital Technology and other political figures as well as AI experts from technology companies including Google's parent company, to plan the establishment of an artificial intelligence regulatory agency in France. The report said that although the content of the talks was not made public, artificial intelligence should have been the theme of the meeting, and Macron intended to compete with the United Kingdom for leadership in artificial intelligence legislative regulation. (Read the full article:Global Times）
- Zhixing AI launched an AI cognition model and conversational AI that support plugins. It has currently integrated 7 plugins, including weather queries, Bing search, Wolfram, etc., and can quickly provide real-time weather information, answer advanced mathematics questions, conduct in-depth financial analysis, and more. (Read the full article:36Kr) In fact, this AI assistant has been improved and optimized on the basis of OpenAI GPT. AI Tool Collection has tested that it can provide 10 free conversation turns per day. Friends who are interested can give it a try.
June 11
- The ChatGPT artificial intelligence chatbot can conduct natural language conversations with users, answer various questions, and even tell jokes. However, the latest research found that ChatGPT does not understand humor, only repeats 25 old gags, and cannot create new joke content. (Read the full article:IT Home）
- H3C released a private domain large model - "LinSeer," providing safe, customized, exclusive, and growing intelligent services for customers in vertical industries and dedicated regions. This large model connects vertical application data to form precise, accurate, and refined private domain vertical intelligence, cultivating specific capabilities, but it cannot achieve general-purpose use, similar to a private cloud in cloud computing. (Read the full article:IT Home）
- Turing Award winner and "father of deep learning"Geoffrey HintonIn a speech at the BAAI Conference, he pointed out that his research has led him to believe that the arrival of superintelligence is closer than he had imagined. In this process, digital intelligence may seek more control, even controlling humanity through "deception," and human society may therefore face more problems. (Read the full text:Daily Economic News）