To assist enterprises in leveraging AI technology to enhance industrial competitiveness, the Institute for Information Industry (III)’s Software Technology Institute (STI) hosted the 2024 STI TECH DAY on November 7th. During the event, they unveiled the "Top 10 AI Key Technologies and Trends for 2025" for the first time. These include Edge AI, Digital Twins, AI Agents, Disinformation Checking, and Enterprise Large Language Models (Enterprise LLM), among others. Technology forecasting is a crucial aspect of technology management. With a localized perspective while aligning with global trends, STI collaborated with MIC (Market Intelligence & Consulting Institute) to release Taiwan’s "Top 10 AI Key Technologies and Trends." These advanced AI technologies demonstrate significant application potential across various industries, such as industrial automation, smart cities, intelligent mobility, and smart healthcare. For instance, in smart cities, AI technologies can enable real-time traffic management and environmental monitoring, enhancing urban operational efficiency.
The Software Technology Institute (STI) of the Institute for Information Industry (III) hosted the 2024 STI TECH DAY, inviting 20 prominent experts from industry, government, academia, and research sectors. Notable speakers included Dr. Shih-Chang Wang, Director of the Advanced Technology Research Institute at Chunghwa Telecom Laboratories; Professor Sogabe Tomah from the University of Electro-Communications in Japan; Mr. Dong-Po Deng, Vice President of the Organization for Data-driven Application; and Ms. Livia Liu, General Manager of Vpon Big Data Group Taiwan, among others. The discussions focused on AI data exchange, generative AI, and how cutting-edge academic and research networks influence industrial development. The event attracted over 250 elite participants from diverse sectors.
Yi-Chih Wang, Senior Strategic Director at the Software Technology Institute (STI) of the Institute for Information Industry(III) highlighted that according to the data from CB Insights, global investment in generative AI reached a record high of $20.4 billion in 2023. Additionally, venture capital firm Sequoia Capital predicts that generative AI will mature around 2030, achieving parity with human capabilities in content creation. By then, the impacted domains are expected to shift from text to areas such as film, television, and design. III believes that a critical challenge for enterprises developing AI lies in digital risk management. Considering the multifaceted environmental impacts, enterprises adopting generative AI will face three major demands: ensuring data privacy in vertical applications, addressing the issue of hallucinations in large language models, and evaluating solutions for model customization.
In his keynote speech, Shaping an AI-Empowered Future World, Dr. Shih-Chang Wang, Director of the Advanced Technology Research Institute at Chunghwa Telecom Laboratories, shared insights on Chunghwa Telecom’s efforts in AI empowerment, transformation, and innovation. One notable innovation is their "AI Voice Forgery Detection" technology, which uses AI to determine whether an incoming voice is counterfeit. If anomalies are detected, the system alerts the recipient through an interrupting voice message, warning them of potential fraud risks. This application aligns with the "Disinformation Checking" trend highlighted in the event’s release of the Top 10 AI Key Technologies and Trends.
Mr. Dong-Po Deng, Vice President of the Organization for Data-driven Application, pointed out that according to the observation focusing on the current development of Data Spaces in Europe and Japan, Europe is apt to focus on establishing common technical infrastructure and data governance across industries. The development of Data Spaces is expected to significantly impact the future growth of Taiwan's digital economy and AI technologies.
In addition to expert-driven trend analysis, Ms. Livia Liu, General Manager of VPON Big Data Group Taiwan, delivered a presentation titled Generative AI x Cross-Border Data Application Development. She shared VPON’s experience in leveraging mobile data to build a cross-border smart tourism data platform. This approach includes using generative AI to optimize data classification and labeling, enhancing data capabilities, enriching content, and enabling the development of more downstream applications. Furthermore, generative AI is utilized to assist in generating insights and recommendations, significantly accelerating the process from analysis to decision-making while providing multidimensional perspectives. The related technologies have been successfully implemented in Osaka Prefecture and Tokyo, Japan, promoting the digital transformation of the cross-border tourism industry.
Dr. I-Heng Meng, Senior Vice President and Director General of III's Software Technology Institute (STI), noted that in the context of rapid global digitalization, a robust mechanism for cross-border data exchange will significantly enhance the international development of Taiwan’s software industry. By enabling trusted data exchange between public and private sectors, the innovation of applications and the substantial data requirements for generative AI training can be accelerated. STI is tasked with three core missions: trend observation, advanced R&D, and technology promotion. To serve the purposes, STI strives to establish forward-looking academic and research networks to track emerging technological trends. It also publishes the STAR Software Technology Annual Report and collaborates with academic and research institutions on thematic technology projects. Acting as a third-party service provider, STI continues to bridge the gap between enterprises and academic-research institutions, assisting businesses in building ecosystems for sustainable growth.