Large Models Bring New Experiences: Are You Willing to Pay for 'Generative AI'?
-
Voiceovers, writing, painting, knowledge dissemination, psychological counseling... As its capabilities continue to improve, generative artificial intelligence (AI) is rapidly integrating into people's lives, playing an increasingly important role.
If generative AI were to introduce itself, its answer might be: Generative AI is a technology capable of producing text, images, sounds, videos, code, and other content based on algorithms, models, and rules. It can simulate human creative thinking, generating logically coherent language, images, audio, and more, enabling interactive dialogue with humans.
What services can generative AI currently provide? Are consumers willing to pay for it?
Large Models Bring New Experiences
At the end of last year, ChatGPT, a chatbot developed by OpenAI, an American AI research company, sparked a wave of enthusiasm for generative AI models. This year, Chinese companies have also launched distinctive generative AI products such as '360 ZhiNao,' 'Dream Diary,' and 'iFlytek Spark.'
A rough estimate shows that there are nearly 200 generative AI products on the market, which can be broadly categorized as follows: interactive chatbots like ChatGPT, Wenxin Yiyan, Tongyi Qianwen, and Baichuan Big Model; intelligent writing tools like Yizhen Miaochuang, Kuaiwen Intelligent Writing, Huolongguo Writing, and Resume Superman; audio and video processing tools like 360 Hongtu, NetEase Tianyin, and SenseTime AI Digital Human; and programming, design, and translation tools like Weiciyun, Alibaba Icons, CodeWhisperer, and Kingsoft PowerWord.
What New Experiences Does Generative AI Offer?
Take the topic 'Should generative AI charge fees?' as an example. If you type this into a traditional search engine, the results would be a list of web links containing keywords like 'generative AI' and 'fees,' but the content may not directly address the question.
However, if you ask this question directly to a generative AI model, you can receive a clear and comprehensive response.
For instance, Alibaba Cloud's 'Tongyi Qianwen' answered: 'This mainly depends on the positioning and usage scenarios of the generative AI product. For basic tools like text or image generators, they can be provided for free. For advanced, commercially valuable products like AI assistants or customer service, charging may be necessary to generate revenue. Additionally, pricing must consider market competition, user demand, and costs.'
Baidu's 'Wenxin Yiyan' responded: 'The answer isn’t a simple yes or no, as it depends on factors like the type, purpose, and cost of the generative AI. Some simple generative AI technologies may be easy to implement and low-cost, possibly not requiring fees. However, more complex technologies may need significant resources and R&D investment, justifying charges. Thus, whether generative AI should charge depends on the specific situation.'
Wei Kai, Deputy Director of the China Academy of Information and Communications Technology's Cloud Computing and Big Data Research Institute, said: 'A standout feature of generative AI driven by large models is its strong adaptability to multiple tasks. This breaks the traditional paradigm where one AI algorithm could only solve a specific problem. Currently, generative AI can serve as a valuable assistant in digital content creation, software development support, customer service, and knowledge management, benefiting professionals in fields like short videos and live-streaming e-commerce. However, these applications are still in their early stages.'
Consumers Are in No Rush to Pay
Not long ago, at the 2023 Tencent Global Digital Ecosystem Conference, Tencent's Hunyuan large model made its official debut. With the integration of this model, many scenarios are expected to undergo significant transformations: After hours-long online meetings, AI technology will be able to quickly and accurately summarize different viewpoints and generate meeting minutes through simple natural language instructions; documents can be automatically formatted with one click, mastering hundreds of Excel formulas and automatically creating charts based on table content; increasingly more advertising creatives generated by large model technology will not only feature industry or regional characteristics with personalized approaches, but also achieve seamless integration of text, images, and videos...
Professor Deng Zhidong from Tsinghua University's AI Research Institute believes that the value of generative AI large models lies in their applications. Only by empowering intelligent economic and social development across diverse practical scenarios can they realize industrial value and achieve their own success.
Interviews revealed that how generative AI functions in production and daily life significantly influences people's willingness to pay.
On one hand, consumers expect tangible results from their payments.
Zhao Shiming, a video creator based in Kunming, tried using generative AI for video production months ago but was unsatisfied with the results. "I mainly wanted AI to assist in creative processes—enriching vocabulary, generating video scripts, and enhancing video quality. However, the output was generic text without distinct features or relevance," he said, emphasizing that product optimization matters more than pricing.
On the other hand, consumers are more willing to pay for high-quality, personalized services.
Wang Suwei, a student in Beijing, noted differences between free and paid versions of ChatGPT: "The paid 4.0 version offers noticeably better database access than the free 3.5 version." She supports tiered pricing models where basic services remain free while advanced features require payment. "If AI can help outline my writing, I'd gladly pay," she added.
For many, supporting the technology's growth is another reason to pay. Ms. Ge from Shenzhen stated: "Payment sustains industry development, protects creators' IP rights, and drives innovation—ultimately improving AI's problem-solving capabilities."
Sustainable Development is Key
Businesses and consumers share a consensus: Any business model should enhance user benefits while ensuring generative AI's healthy, sustainable growth.
Ms. Ni, a UI designer in Hangzhou, reported significant efficiency gains from a paid generative AI tool: "While outputs still need refinement, it expands creative possibilities." She hopes for more powerful features in future products.
According to Zhang Xiangzheng, president of 360 Zhinao, their AI services now cover over 100 life and work scenarios (e.g., smart creation, office, socializing) for individual users, with enterprise solutions implemented across nearly 20 industries including finance, healthcare, and education.
"AI-related services require significant investments in data and computing power, making the commercialization of large models an unavoidable proposition. Traditional internet business models typically include advertising and premium membership services, generating revenue through application scenarios. In the early stages of generative AI commercialization, diversified payment models may help chart the course for industry development. We believe that generative AI services should be tiered: basic internet services and universal needs should remain free, while personalized 'premium AI services' such as AI digital humans could charge fees," said Zhang Xiangzheng. He added that 360 will also launch higher-quality generative AI products and services to capture more business opportunities by addressing user needs.
McKinsey & Company projected in a report titled Generative AI in China: $2 Trillion in Economic Value that generative AI could contribute approximately $7 trillion to the global economy, with China accounting for about $2 trillion—nearly one-third of the total.
Wei Kai explained that the current development and operational costs of generative AI are high, involving investments in software development, training computing power, as well as maintenance and inference computing power. Particularly during the service phase, the computing costs for generative AI far exceed those of web browsing or search engines. Therefore, it is not only a reasonable business decision for companies to charge appropriately for large model services but also a pathway to promoting the sustainable development of the generative AI industry. "Specifically, the conditions and methods of charging will depend on market feedback," Wei Kai noted.
It is worth noting that the Interim Measures for the Management of Generative Artificial Intelligence Services, issued by the Cyberspace Administration of China and other departments, took effect on August 15, providing institutional safeguards for balancing development and security. Many industry experts pointed out that generative AI may pose risks such as harmful speech, misinformation, privacy breaches, and biases in practice. Security is the foundation for generative AI to gain market acceptance. Establishing a healthy, secure, and sustainable business model—where valuable innovations and services receive fair returns—is a practical necessity for the high-quality development of the generative AI industry.