LABScon24 Replay | The Real AI Race: Disinformation in the Taiwanese Election | Martin Wendiggensen

AI disinformation, deep fakes, and bot manipulation are often portrayed as the greatest threats to elections. However, Taiwan’s 2024 election defied these expectations, with no significant AI disinformation campaigns from China, despite high-level government focus.

In this talk, Martin Wendiggensen shares his investigation into the election, analyzing vast amounts of footage, articles, and social media content through a multi-modal AI pipeline. Surprisingly, the results indicated that AI-generated content had minimal impact due to lack of engagement.

Instead, the real influence came from Taiwanese billionaires with financial ties to China, who invested in or created local media outlets. These outlets pushed narratives favoring Beijing-friendly parties, influencing millions of viewers. While these parties did not win the presidency, they managed to secure parliamentary control and are set to shape Taiwan’s future policies.

Martin discusses how open-source AI tools were used in the study, offering critical lessons on countering disinformation and protecting election integrity.

Read the full blog post here: https://s1.ai/LC24-MW