1. Categories of AI Applications: How AI Is Used Across Industries
According to the report “Artificial Intelligence in Taiwan: Opportunities and Challenges for Industrial Transformation”, current AI applications fall into two main categories:
- Basic cross-industry needs
- Shared needs within specific industries
Basic cross-industry applications include facial recognition, emotion detection, speech recognition, natural language understanding, object recognition, behavior analysis, and scene detection.
Industry-specific applications include defect detection in manufacturing, risk prediction in finance, sales forecasting in retail, route planning in logistics, and sunlight/watering control in agriculture.
Facial recognition is now widely used. This technology identifies facial features by analyzing various points on the face and compares the results with a facial database to verify identity—enabling functions like facial payment.
When a person displays emotions, these facial feature points shift. AI can track these changes to monitor emotional states. Meanwhile, object recognition and scene detection are often applied in transportation, for example, to detect vehicle distance, license plates, pedestrians, and traffic signs.
2. AI in Taiwan’s Traditional Industries: Taxis and Textiles
AI has entered traditional industries in Taiwan, including the taxi and textile sectors.
For example, Taiwan Taxi partnered with Acer to apply AI and big data for:
- Hotspot analysis
- Demand forecasting
- Smart recommendations
This helps drivers better understand where and when passengers need rides, optimizing their time and routes.
- Hotspot Analysis identifies areas with the highest ride request volumes based on past data. It incorporates time, day, location, and external factors like weather, holidays, and events to recommend active areas to drivers.
- Demand Forecasting uses machine learning and statistical models to predict demand in various regions and time slots, based on real-time and external conditions.
- Smart Recommendations combine hotspot and demand analysis with the driver’s current location to suggest where they should go next—ensuring real-time balance between supply and demand.
In the textile industry, Ho-Ming Textile in Tainan, with over 40 years of history, introduced AI to improve the manual and time-consuming fabric identification process. Previously, it could take 1 to 2 months to manually match fabric samples.
Since 2017, the company has collaborated with Google’s machine learning and cloud platforms to digitize its vast library of fabric samples and build a fabric recognition model. Designers can now scan fabric samples via mobile app to identify them instantly. This AI tool shortened the design process from 1.5–3 months to just 2–3 days.
3. AI in Everyday Technology
FaceHeart, led by Professor Wu Bing-Fei from National Chiao Tung University, uses AI image analysis to detect subtle changes in facial microvessels via regular cameras or smartphone lenses. This enables non-contact measurement of heart rate and blood pressure, with accuracy comparable to medical devices—providing users with quick, accessible health insights.
4. AI as a Marketing Tool
AI has become increasingly embedded in marketing, most notably in chatbots, which automate customer engagement and content delivery. AI is also widely used for consumer behavior prediction and personalized recommendations based on purchase history, helping brands achieve more precise targeting.
For example, Rosette.ai provides recommendation systems for e-commerce platforms. By analyzing data such as customer lists and product catalogs, it generates personalized promotional suggestions using AI algorithms.
According to its founder, returning customers are often the most valuable. However, many e-commerce platforms suffer from poorly designed interfaces and high development costs, limiting conversion rates. Rosette.ai addresses this by offering 30+ plug-and-play API modules, designed to match various consumer journeys and business KPIs. Customers using the system have seen an average 15% increase in click-through rates and 25% lift in add-to-cart rates.
In addition to personalization, AI also streamlines content production. GliaStudio, an AI-powered video creation platform, automatically transforms articles into video content. By analyzing text using natural language processing (NLP), GliaStudio identifies key themes, segments the content, finds relevant visuals from its media library, and compiles them into videos—significantly reducing production time.
Another crucial area is marketing performance tracking. Platforms like Smart Engage from adGeek use AI to collect, unify, and analyze both real-time and historical data from multiple channels. This helps marketers fine-tune their strategies and deliver more accurate consumer targeting.
5. AI’s Hidden Risks and the Growing Demand for Talent
Although AI reduces labor demands in some areas, it also generates new types of job opportunities.
Michael Kung, Senior Partner at Gartner Greater China, projected that by 2023, 75% of large enterprises would hire “AI auditors” to monitor ethics, customer privacy, and compliance—reducing corporate risk. With increased digital interconnectivity, cyberattacks have also become more frequent, posing threats to data integrity and accuracy.
Future AI threats may include:
- Contaminating training data
- Stealing AI models
- Injecting malicious code at the data source to manipulate model outcomes
Such risks highlight the importance of data accuracy and oversight, as AI lacks common sense and ethical judgment. The quality of input data directly influences AI output.
This is why Alphabet (Google’s parent company) warned investors that AI may lead to legal and ethical conflicts that could affect profits. For instance, in 2015, Google Photos sparked backlash when it mistakenly labeled Black users as “gorillas,” revealing the need for strong ethical supervision in AI.
In terms of workforce, a report by Taiwan’s Industrial Development Bureau (2019–2021) identified 12 critical AI roles, including:
- Machine Learning Engineers
- Algorithm Engineers
- Software Engineers
- Cloud System Engineers
- QA/Test Engineers
- UI/UX Designers
- Data Scientists
- Data Engineers
- Data Analysts
- Data Labeling Specialists
- AI Project Managers
- Product Managers/FAEs
Conclusion
The rapid evolution of technology may eliminate some roles while creating others. But one thing remains unchanged: the interdependence of humans and technology. As we embrace the innovation and impact that AI can bring, we must also reflect on how to manage the relationship between humans and machines with care and foresight.
