The Legislative Yuan’s full-house committee reviewed President Lai Ching-te’s impeachment case. Kuomintang legislator Ko Ju-yun criticized that the Lai government’s progress in fulfilling campaign promises, the energy transition, national security, economic and trade diplomacy, traffic safety, and AI equality has not met expectations.
In his remarks, Ko Ju-yun said that during Lai Ching-te’s campaign he put forward 227 policy pledges, but so far only 2 have been actually implemented, while most of the rest are still at the stage of not having started. He criticized that if the pledges cannot effectively be translated into action, it amounts to a political failure to keep faith with the people, and also constitutes a political responsibility for administrative inaction.
From energy policy to national security, Ko Ju-yun lists Lai Ching-te’s five major governance problems
Ko Ju-yun first addressed energy policy. He said Lai had promised to build a net-zero innovation technology platform, but Taiwan still relies heavily on thermal power generation, with an overly high share of thermal power. Coal-fired and gas-fired power remain the main sources of electricity.
He criticized that while the government maintains the non-nuclear homeland route, it has allowed air pollution and energy risks to be borne by the people, questioning whether such an energy policy matches the people’s health and the industry’s needs for stability. Ko Ju-yun also argued that many countries internationally have re-evaluated nuclear power as a clean energy option, and Taiwan should not completely exclude nuclear power from energy discussions.
The second issue is national security and communication resilience. Ko Ju-yun said Taiwan heavily relies on submarine cables for its external internet connectivity. If submarine cables are damaged or if there is a conflict risk, Taiwan could face a disruption of external communications, or even become an information island.
He believed the government should more proactively introduce diversified satellite communication services and incorporate satellite networks as part of national resilience. However, in his view, while the Lai administration has promised to strengthen defense and national security capabilities, there is still no concrete progress at present in diversified satellite backups and national communications resilience.
Ko Ju-yun then mentioned economic and trade diplomacy. He said Lai Ching-te’s campaign pledges included promoting Taiwan’s accession to regional economic cooperation mechanisms such as the CPTPP, but the progress and specific timeline remain unclear.
The fourth issue is traffic safety. Ko Ju-yun said Lai had promised to improve road traffic safety and set a goal of “zero traffic deaths by 2040,” but Taiwan’s current laws and regulations related to vehicles and autonomous driving still lag behind international standards. He cited the example of Tesla’s FSD being approved for use in some countries, and criticized Taiwan’s slow pace in opening regulations for advanced driver-assistance and autonomous driving technologies, which results in “Taiwan can make chips, but cannot allow those advanced technologies to be used on Taiwan’s roads.”
Ko Ju-yun also mentioned the “three shifts to care for the sick” issue. He said the Ministry of Health and Welfare originally took a conservative stance toward promoting the related system and even said it would require two years before implementation. But after Lai Ching-te and the minister had brief talks, the policy direction quickly reversed.
Seventy percent of people using free AI also becomes a reason for the impeachment
However, throughout the entire speech, what drew the most discussion was not traditional political power struggle, but Ko Ju-yun listing “more than 70% of Taiwan residents can only use free versions of AI tools” as one of the reasons for the president’s failure in governing, and even further linking it to “AI poverty” and “AI equality.”
From the perspective of public policy, AI equality is indeed an important issue. Generative AI is rapidly entering education, the workplace, research, software development, content production, and industry decision-making processes. Whether people can obtain high-quality AI tools in the future, and whether they have AI literacy, may truly affect individuals’ competitiveness and social mobility. If only high-income people, companies, or certain schools can access advanced AI tools, the digital divide may further expand into a productivity divide.
Ko Ju-yun’s argument seems to interpret AI equality as “making it affordable for more people to subscribe to advanced AI.” But AI equality should not only mean that the government pays for people’s paid accounts. Real and effective AI equality should include public educational resources, AI literacy in schools, the ability to use open-source models, Chinese data and local application scenarios, introducing AI into public services, and preventing disadvantaged groups from being excluded by algorithms. If it is only turning the prices of paid AI tools into a political accusation, it will instead narrow AI policy into a “subscription-based subsidy” problem.
An a16z report from earlier this year from Silicon Valley venture capital already pointed out that users’ main needs for AI models (taking ChatGPT as an example) are concentrated on daily productivity:
Writing and editing: 28.1%
Practical advice (life, health, learning, etc.): 28.3%
Information lookup: 21.3%
By comparison, computer programming is only 4.2%, and high-value scenarios such as data analysis and math account for even less. Engineers or professional users can gain several times the productivity from AI, so they are willing to pay $20 to $200 per month. But for ordinary users who merely look up information, write emails, or ask questions, free services are already sufficient.
(What does the ChatGPT advertisement look like? When 90% of users aren’t willing to pay, AI is moving toward “a small number of paid users, while the many watch ads.”)
Especially, “free-version AI” itself does not mean it cannot be used, is unfair, or is behind. Many free AI tools can already handle search, summarization, translation, writing assistance, code explanations, and learning tasks. For the general public, the problem is often not “whether there is a paid model,” but whether they know how to use it, how to recognize hallucinations, how to integrate AI into their work workflows, and whether schools and workplaces provide basic training.
Ko Ju-yun also said in his remarks that as long as there are great AI tools that can help with searching, young people can know that “the Earth isn’t round; the Earth is oval.” The intended meaning may have been to highlight how AI can bridge educational and knowledge gaps, but placed within an impeachment speech, it makes the entire AI equality argument look more like political rhetoric rather than a complete policy proposal.
AI tools can indeed help people look up knowledge, but the core of AI equality is not only “searching for answers.” The key is how to give people the ability to judge answers, verify sources, use tools, and avoid misinformation. If AI is treated as an all-purpose search tool, or even used to package impeachment reasons, it may end up overlooking the part of generative AI that most needs to be taught: it is not an authority-answer machine, but a productivity tool that must be checked and verified.
This article, Legislative Yuan reviews Lai Ching-te impeachment case; Ko Ju-yun lists five governance problems: 70% of people using free AI also becomes a reason for impeachment, was first published on Chain News ABMedia.
