(852) 91354113
Since OpenAI launched ChatGPT late last year, both Western and Chinese tech giants have raced to develop their own AI chatbots, eager to capitalize on the AI revolution. While larger companies can leverage ChatGPT in various ways to enhance their operations, smaller enterprises face challenges due to their limited technical capabilities.
Derek Ng Tak-chuen, director of Hong Kong-based startup Parami, which specializes in AI chatbots and solutions, suggests that smaller firms might benefit more from partnering with AI specialists rather than attempting to develop and integrate ChatGPT independently. Ng notes that some companies are hesitant to adopt ChatGPT due to concerns about security risks and the potential for the chatbot to provide incorrect answers or solutions.
Ng explains that integrating ChatGPT into a company's operations essentially means delegating part of the business process to the AI tool. However, this transfer can be problematic if the AI isn't reliable in specific scenarios. He points out that creating ChatGPT-like products is relatively straightforward for internet giants due to their vast data resources, but emphasizes that data quality is just as important as data volume in developing effective AI models.
Earlier this year, Chinese search engine giant Baidu introduced its own ChatGPT-like product, Ernie Bot. Despite its high expectations, the launch fell short, with CEO Robin Li Yanhong admitting the chatbot was "not perfect." Ng remarks, "We've analyzed Baidu’s language model and found its quality lacking." He attributes ChatGPT's superior performance to the extensive efforts put into maintaining data quality.
Despite some criticism—such as instances where ChatGPT struggled with complex math questions—businesses and developers continue to push forward. Chinese technology giants Alibaba and Tencent are also planning to release their own ChatGPT-style products. Additionally, telecom companies like China Telecom and China Unicom are developing their own AI models.
Ng suggests that collaboration between telecoms and tech giants could increase the chances of success. For instance, if China Mobile combines its audio data with Baidu’s text data, it could potentially develop a multimedia GPT. However, he also notes that China's internet giants might face regulatory pressures due to their vast data access.
Ng believes that the key to successfully commercializing AI tools like ChatGPT lies in developing vertical applications tailored to specific business needs. While ChatGPT can enhance various functions such as customer service, content creation, and coding, the AI business model remains a work in progress. Ng emphasizes that for AI tools to be widely adopted, there must be clear, measurable indicators of their effectiveness—an aspect currently lacking in the AI field.
