Chinese companies are doubling down on open-source artificial intelligence-powered models as part of a broader push to bring fast-evolving AI technology to more businesses and developers worldwide, and bolster its application in a diverse range of fields.

Experts said the open-source approach will lower the threshold for the development and application of AI, greatly reduce computing power costs, and foster the sharing of AI tech around the world, as well as boost collaboration and innovation.

The recent progress in open-source large language models has showcased China”s growing technological prowess and open attitude in the AI domain, given that Chinese AI startup DeepSeek’s open-source models have taken the world by surprise, they added.

Tech heavyweight Alibaba Group has stepped up efforts to enable broad access to its AI technology and innovations by releasing large language models from its Qwen family as open-source, and boasting China’s largest AI open-source community platform, ModelScope.

The company has made more than 200 generative AI models open-source in recent years. The models have multimodal capacities and can process and generate various types of content, covering text, images, audio and video.

ModelScope, which was launched in November 2022, hosts over 70,000 open-source models, and the user base has expanded from 1 million in April 2023 to 16 million as of June 30, serving 16 million developers from 36 countries around the world.

It supports developers in experiencing, downloading, fine-tuning, training and deploying models. Various types of open-source AI models have been included in the community.

“We aim to simplify and reduce the cost of developing, customizing and deploying AI models for developers and corporations, thereby enabling the creation of revolutionary AI applications that have a positive impact on society,” said Zhou Jingren, chief technology officer at Alibaba Cloud Intelligence, emphasizing they are committed to making AI models more accessible and easier to use.

Baidu Inc has recently open-sourced its multimodal LLM Ernie 4.5 series, consisting of 10 distinct variants. The model family includes mixture-of-experts (MoE) models with 47 billion and 3 billion parameters, the largest model having 424 billion parameters, alongside a 0.3 billion dense model. The Ernie 4.5, launched in March, is Baidu’s multimodal foundational model.

The company said the MoE architecture has the advantages of enhanced multimodal understanding and improved performance on text-related tasks. All models are trained with optimal efficiency using the PaddlePaddle deep learning framework, which enables highper-formance inference and streamlined deployment.

Experimental results show that the models achieve state-of-the-art performance across multiple text and multimodal benchmarks, particularly in instruction following, knowledge memorization, visual understanding and multimodal reasoning.

Zhu Keli, founding director of the China Institute of New Economy, said the open-source approach adopted by a string of Chinese AI companies will lower the technical threshold, speed up the popularization of AI tech across various sectors including automobiles, manufacturing, finance and education, and allow more enterprises and developers to participate in AI research and development.

Zhu believes technological innovation is unstoppable, and international cooperation serves as an important way to promote the development of AI tech, adding that China’s open and inclusive attitude helps promote the advancement of the global AI industry.

“Open source will allow resource-constrained startups, small businesses and entrepreneurial developers to access cutting-edge AI tech and build their own models more cost-effectively,” said Pan Helin, a member of the Expert Committee for Information and Communication Economy, which is part of the Ministry of Industry and Information Technology.

Pan said it will accelerate AI technological advancements and breakthroughs by enabling global developers to create customized industry-specific models, and foster a more competitive and diverse AI ecosystem.

Chinese AI companies have the ability to take the lead in global AI innovation, as they have sought an alternative AI development approach that emphasizes efficiency and open-source collaboration — which is different from their US counterparts — while reshaping the global AI landscape, Pan added.

The market size of the nation’s AI sector will reach 1.73 trillion yuan ($241.2 billion) by 2035, accounting for 30.6 percent of the global total, said market research firm CCID Consulting.

Hello, I’m real-life Amy. I’m an instructional designer at Hong Kong UST. Hello, I am AI Amy. Can you see the difference? From now on, this will be narrated by AI. I am excited to share how generative AI can unlock the potential of repurposing online recordings for enhanced blended learning. Educational videos are now essential resources in higher education, serving as vital content delivery tools in flipped, blended and online classes. However, producing high-quality video recordings can be challenging. 

In the past year, the global education community has faced significant challenges due to the rapid shift toward digital learning. Traditional face-to-face classrooms have transformed into online and hybrid formats, necessitating quick adaptation from faculty and teaching staff. Many have turned to platforms like Zoom, resulting in a wealth of recorded content. Yet, while these recordings often serve as backups, they have the potential to be transformed into effective educational tools. Today’s students are no longer confined to textbooks and lecture notes. They increasingly turn to videos to learn software, solve problems and acquire new skills. As educators, we may hesitate, wondering if we need to master video editing or recording techniques.

Achieving clarity and engagement in a single take is challenging. This is where generative AI steps in to simplify and streamline the content enhancement process. By leveraging generative AI technologies, we can repurpose our recordings for blended learning environments, making them more accessible and engaging. 

Importantly, this innovative approach allows us to enhance materials for future iterations without the need for constant rerecording. By analysing student feedback and learning patterns, we can refine our videos for subsequent runs, ensuring our online content continually evolves to meet learners’ needs. This adaptive strategy not only boosts student engagement but also maximises the effectiveness of our teaching strategies.

Generative AI provides practical solutions to enhance your video development process. Here’s how: 

Step one, transcription. Utilise the speech to text capability of generative AI to convert audio extracted from lectures, whether face-to-face or Zoom, into written transcripts. This feature allows for easier review and refinement of your materials. By automating transcription, you save valuable time and can focus on enhancing the clarity of your content.

Step two, review and refine. Once transcription is complete, you can edit the script to improve coherence and align the content with desired learning outcomes. This step is crucial for ensuring that your message resonates with students and effectively conveys the intended concepts. 

Step three, voice cloning and AI mirror avatar. Generative AI also enables you to create voice clones and AI avatars, eliminating the need for extensive re-recording. This capability allows you to deliver engaging, error-free speech, facilitating a smoother learning experience for students as they follow along and grasp complex concepts more easily. 

Let’s explore a scenario where an instructor has recorded content and identified opportunities for refinement during the review process. With the assistance of generative AI, the instructor can streamline content refinement in three simple steps, prioritising substance over technical challenges. Transcribe. Upload your recorded audio from a face-to-face lecture or Zoom session to a speech to text tool. The transcription process is quick and efficient, allowing you to add the text to your document for further refinement.

Review and edit. With the transcription in hand, edit the script to enhance clarity and coherence, aligning it with your learning objectives. Clone your voice and your own AI avatar. Utilise digital clone technology to create a digital version of your voice and an AI avatar. This process is straightforward and involves either recording new audio video or uploading an existing sample.

Once everything is ready, we can thoughtfully integrate the digital version of your voice and your AI avatar into the PowerPoint slide deck, allowing for easy addition or editing of new content without affecting the entire video rerecording. Please pause the video and scan the QR code to see how this works in the final output of the educational video. 

By leveraging generative AI technologies, these tools provide time-saving benefits for upcycling recorded content. Instructors can effectively repackage their knowledge assets and implement a blended learning approach. While AI tools offer significant advantages, they are designed to assist, not replace, educators’ expertise. Your role in reviewing and refining content is essential to ensure accuracy and alignment with instructional goals. For educational videos to truly engage learners, technical skills must be paired with sound pedagogical principles to maximise learning effectiveness. Integrating generative AI into video production not only saves time but also reduces the need for extensive space and equipment set-up. Most importantly, it enhances the overall learning experience by utilising the expertise of content specialists in script development. By balancing technology with human involvement, we can create more effective and engaging educational resources for our students.

Amy Chong is an instructional designer at Hong Kong University of Science and Technology.

If you would like advice and insight from academics and university staff delivered direct to your inbox each week, sign up for the Campus newsletter.

Sign up for Chalkbeat’s free weekly newsletter to keep up with how education is changing across the U.S.

More than 400,000 K-12 educators across the country will get free training in AI through a $23 million partnership between a major teachers union and leading tech companies that is designed to close gaps in the use of technology and provide a national model for AI-integrated curriculum.

The new National Academy for AI Instruction will be based in the downtown Manhattan headquarters of the United Federation of Teachers, the New York City affiliate of the American Federation of Teachers, and provide workshops, online courses, and hands-on training sessions. This hub-based model of teacher training was inspired by work of unions like the United Brotherhood of Carpenters that have created similar training centers with industry partners, according to AFT President Randi Weingarten.

“Teachers are facing huge challenges, which include navigating AI wisely, ethically and safely,” Weingarten said at a press conference Tuesday announcing the initiative. “The question was whether we would be chasing it or whether we would be trying to harness it.”

The initiative involves the AFT, UFT, OpenAI, Microsoft, and Anthropic.

The Trump administration has encouraged AI integration in the classroom. More than 50 companies have signed onto a White House pledge to provide grants, education materials, and technology to invest in AI education.

In the wake of federal funding cuts to public education and the impact of Trump’s sweeping tax and policy bill on schools, Weingarten sees this partnership with private tech companies as a crucial investment in teacher preparation.

“We are actually ensuring that kids have, that teachers have, what they need to deal with the economy of today and tomorrow,” Weingarten said.

The academy will be based in a city where the school system initially banned the use of AI in the classroom, claiming it would interfere with the development of critical thinking skills. A few months later, then-New York City schools Chancellor David Banks did an about-face, pledging to help schools smartly incorporate the technology. He said New York City schools would embrace the potential of AI to drive individualized learning. But concrete plans have been limited.

The AFT, meanwhile, has tried to position itself as a leader in the field. Last year, the union released its own guidelines for AI use in the classroom and funded pilot programs around the country.

Vincent Plato, New York City Public Schools K-8 educator and UFT Teacher Center director, said the advent of AI reminds him of when teachers first started using word processors.

“We are watching educators transform the way people use technology for work in real time, but with AI it’s on another unbelievable level because it’s just so much more powerful,” he said in a press release announcing the new partnership. “It can be a thought partner when they’re working by themselves, whether that’s late-night lesson planning, looking at student data or filing any types of reports — a tool that’s going to be transformative for teachers and students alike.”

Teachers who frequently use AI tools report saving 5.9 hours a week, according to a national survey conducted by the Walton Family Foundation in cooperation with Gallup. These tools are most likely to be used to support instructional planning, such as creating worksheets or modifying material to meet students’ needs. Half of the teachers surveyed stated that they believe AI will reduce teacher workloads.

“Teachers are not only gaining back valuable time, they are also reporting that AI is helping to strengthen the quality of their work,” Stephanie Marken, senior partner for U.S. research at Gallup, said in a press release. “However, a clear gap in AI adoption remains. Schools need to provide the tools, training, and support to make effective AI use possible for every teacher.”

While nearly half of school districts surveyed by the research corporation RAND have reported training teachers in utilizing AI-powered tools by fall 2024, high-poverty districts are still lagging behind their low poverty counterparts. District leaders across the nation report a scarcity of external experts and resources to provide quality AI training to teachers.

OpenAI, a founding partner of the National Academy for AI Instruction, will contribute $10 million over the next five years. The tech company will provide educators and course developers with technical support to integrate AI into classrooms as well as software applications to build custom, classroom-specific tools.

Tech companies would benefit from this partnership by “co-creating” and improving their products based on feedback and insights from educators, said Gerry Petrella, Microsoft general manager, U.S. public policy, who hopes the initiative will align the needs of educators with the work of developers.

In a sense, the teachers are training AI products just as much as they are being trained, according to Kathleen Day, a lecturer at Johns Hopkins Carey Business School. Day emphasized that through this partnership, AI companies would gain access to constant input from educators so they could continually strengthen their models and products.

“Who’s training who?” Day said. “They’re basically saying, we’ll show you how this technology works, and you tell us how you would use it. When you tell us how you would use it, that is a wealth of information.”

Many educators and policymakers are also concerned that introducing AI into the classroom could endanger student data and privacy. Racial bias in grading could also be reinforced by AI programs, according to research by The Learning Agency.

Additionally, Trevor Griffey, a lecturer in labor studies at the University of California Los Angeles, warned the New York Times that tech firms could use these deals to market AI tools to students and expand their customer base.

This initiative to expand AI access and training for educators was likened to New Deal efforts in the 1930s to expand equal access to electricity by Chris Lehane, OpenAI’s chief global affairs officer. By working with teachers and expanding AI training, Lehane hopes the initiative will “democratize” access to AI.

“There’s no better place to do that work than in the classroom,” he said at the Tuesday press conference.

Chalkbeat New York bureau chief Amy Zimmer contributed to this report.

Norah Rami is a Dow Jones education reporting intern on Chalkbeat’s national desk. Reach Norah at nrami@chalkbeat.org.