Taiwan's opposition KMT is turning the U.S. accusations against Chinese AI firms into a political weapon, which is bad news for China's tech image and awkward for the KMT's own pro-China base.

• KMT "tech lawmaker" Ko Ju-chun, fresh from a U.S. trip, says China's AI lead is fake: rebranded copies that secretly route through American models.
• He points to Anthropic's move last week naming seven Chinese firms — including Alibaba, DeepSeek, Moonshot (Kimi), Xiaomi and SenseTime — for industrial-scale copying of Claude, with Kimi caught forwarding 23 million user queries to Claude in three months.
• He ties this to shaky Huawei chip yields and calls Xi Jinping's planned U.S. visit a sign of weakness.
• Ko pitches a "new-generation KMT" as the true anti-communist party and wants Taiwan-U.S. ties deepened through chips, drones and AI — while also blasting the ruling DPP over a Changhua vote-buying probe.
• The post backfired online: Chinese nationalist commenters mocked him for "believing whatever Americans say," and deep-blue KMT supporters accused him of sounding like a DPP lawmaker.

Elsewhere in Taiwan: Taoyuan's Longtan science park expansion cleared national review, promising big semiconductor output and 4,500 jobs; Changhua's council speaker and his parents face detention over alleged vote-buying; and Taipei's city council froze funding for the Shanghai "Twin City Forum," calling its agreements "zombie MOUs."

Outlook: Expect the KMT to keep testing a harder anti-Beijing line ahead of the November 2026 local elections, with its own base split on how far to go.