Data quality key to good AI-generated content

The rapid rise of artificial intelligence-generated content (AIGC) is transforming the digital economy in China as well as the rest of the world. AIGC, fueled by breakthroughs in AI technologies, such as generative models, natural language processing and deep learning, is changing the way content is produced and consumed.

In China, AIGC is expanding at an unprecedented rate, driven by both government and private sector initiatives. According to a report of the China Academy of Information and Communications Technology, the market value of China's digital economy, bolstered in part by the contributions of AIGC, is likely to exceed 60 trillion yuan ($8.23 trillion) by the end of this year.

At the forefront of this revolution, Chinese tech giants such as Baidu, Alibaba and Tencent are investing heavily in AIGC applications for sectors like e-commerce, media and education. The sector's rapid growth is driven by the increasing demand for automated content creation in marketing, entertainment and customer service. But such growth comes with challenges — particularly of content authenticity, intellectual property rights, and the need for robust regulatory frameworks to address these issues.

A key challenge in AIGC's growth is the quality of data that powers these systems. While the algorithms and computing power behind AIGC are impressive, the effectiveness of these systems is determined by the data they process. This is where data quality management (DQM) becomes crucial. Low-quality data can lead to inaccurate outcomes, which can be particularly problematic in sensitive sectors such as healthcare and social services. A well-known case highlighting the dangers of poor data quality is Google's Flu Trends model. In February 2013, due to data issues, it predicted more than double the proportion of doctor visits for influenza-like illness compared with the official estimates.

Data quality issues in AIGC are very important. AI systems sometimes generate "hallucinations" — false or fabricated content — raising concerns about misinformation. A 2023 report by OpenAI found that their AI text detector correctly identified 26 percent of AI-generated text as "likely AI-written".

Furthermore, Deloitte's 2024 report highlighted that over 50 percent of organizations reported facing significant challenges with data quality in their ESG reporting, which can impact the reliability of data used in decision-making, and potentially contribute to issues like misinformation and the spread of fake news.

In China, AI models also face such challenges due to a lack of high-quality corpus data. Problems related to data diversity and labeling errors persist despite the implementation of the Data Security Law.

In economics, the "invisible hand" refers to unseen forces that guide the free market. However, in the realm of AIGC, the "invisible hand" may represent hidden data issues that can compromise the integrity of the entire system. For this reason, it is crucial for both companies and regulators to verify data outcomes with precision, ensuring that all the data fueling AIGC systems are of the highest quality.

Addressing these challenges is essential to unleash the full potential of AIGC. The importance of DQM cannot be overstated. In China's rapidly growing digital economy, the role of DQM is key to fostering trust in AIGC technologies. High-quality data lay the foundation for reliable insights and better decision-making, essential for the adoption of AIGC across various industries.

To ensure the high quality of data, the authorities have to adopt a multifaceted approach. First, the government should play a leading role in this process, by establishing clear, standardized data regulations, so as to help create a unified framework for DQM and ensure that all stakeholders operate on a level playing field.

Another crucial aspect is the development of a data-centric culture within organizations. Companies must view ensuring data quality as a shared responsibility, with employees trained to understand its importance and equipped with the skills to effectively manage it. By fostering such a culture, businesses can reduce errors and biases in data, which in turn will improve the performance of AIGC systems.

Collaboration between industry and academia, too, is important for advancing DQM in the AIGC domain. Research institutions can develop advanced algorithms to assess and improve data quality, while businesses can provide real-world data and use cases to test these methods. Such partnerships will drive innovation and ensure AIGC technologies are effective, ethical and reliable.

Moreover, the establishment of data markets could play a pivotal role in addressing data scarcity, particularly in specialized fields such as personalized medicines and services. Such markets will enable data sellers to provide valuable insights. But the challenge for buyers is to select the most relevant data points from an often-overwhelming array of options.

For China to fully capitalize on the potential of AIGC and lead the global digital economy, it must prioritize excellence in DQM. By establishing robust data standards, investing in infrastructure, promoting a data-driven culture, and fostering industry-academia collaboration, China can ensure that AIGC technologies are built on solid, reliable data.

This will give China a competitive edge in the global marketplace. As China continues to develop AIGC, the pursuit for high-quality data will be a defining factor in shaping the future of the digital economy and society.

The author is an associate professor at the School of Economics and Management, Chongqing University of Posts and Telecommunications.

The views don't necessarily reflect those of China Daily.