Is AI-Generated Content Legal in Taiwan? What the Law Actually Says

If you're generating AI images or video in or about Taiwan, it's worth knowing where the actual legal lines sit rather than guessing. The short version: Taiwan doesn't have one single law that regulates "AI-generated content" as its own category. What it does have is a real, specific criminal law against making fabricated sexual images of another identifiable person without consent — added to the Criminal Code back in 2023, well before most countries had anything comparable — plus a brand-new AI Basic Act that sets national principles rather than content rules, and a narrower rule requiring online ad platforms to disclose when an advertisement uses AI-generated or deepfake material. None of this is legal advice — it's a plain-language summary of publicly available Taiwanese law as best we can determine it, and if you're doing anything beyond casual personal use, you should talk to a lawyer licensed in Taiwan about your specific situation.

Generate fictional AI content

Taiwan doesn't have a dedicated "AI-generated content" law — it has a dedicated deepfake-sexual-image law

Unlike a general content-restriction statute aimed at AI as a category, Taiwan's most directly relevant law targets a specific harm: fabricated sexual imagery of a real person, however it was made. That's a meaningfully narrower and more specific target than "AI content" broadly, which means plenty of AI generation — fictional characters, non-sexual content, content that doesn't depict any real identifiable individual — simply isn't what this particular law is aimed at.

Taiwan's Criminal Code Article 319-4: the real deepfake-image law

In January 2023, Taiwan's legislature passed an amendment adding a new chapter to the Criminal Code, "Offenses Against Sexual Privacy and Fabricated Sexual Images" (妨害性隱私及不實性影像罪), with Article 319-4 at its center. As published in Taiwan's official Laws & Regulations Database (全國法規資料庫), the article criminalizes creating a fabricated sexual image of another person through computer synthesis or similar technology with intent to distribute, broadcast, deliver, or publicly display it — punishable by up to five years imprisonment, or up to seven years if done for profit. The law applies to fabricated sexual images of identifiable real people generally; it isn't specific to any one category of victim, and it was reportedly motivated by real deepfake-pornography cases involving Taiwanese public figures.

Taiwan just passed a general AI Basic Act, not a content-restriction law

Taiwan's Artificial Intelligence Basic Act (人工智慧基本法) was promulgated on January 14, 2026, giving the country its first dedicated AI statute. Based on available legal analysis, it functions as a framework law: it sets guiding principles, prohibits AI applications that infringe on personal safety, freedom, property, or privacy or that threaten social order or national security, and lays groundwork for future AI policy — but it doesn't itself impose direct compliance obligations on private users or platforms the way a content-restriction statute would. It's a useful signal of where Taiwanese AI policy is heading, not a rulebook for what a specific generated image is or isn't allowed to depict.

A narrower rule already requires AI-content disclosure in online advertising

Separately, Taiwan's Fraud Crime Hazard Prevention Act (詐欺犯罪危害防制條例), specifically Article 31, Paragraph 1, Subparagraph 4, took effect in mid-2024 and requires online advertising platforms to disclose whether an ad uses AI-generated or deepfake technology. It was written to address AI-driven scam ads, not adult content specifically, but it's a concrete example of Taiwan already legislating disclosure obligations for AI-made material in at least one context — a pattern that could plausibly expand.

What this actually means if you generate fictional AI content

For entirely fictional AI characters that don't depict, name, or closely imitate any real identifiable person, Taiwan's highest-profile AI law — the Article 319-4 deepfake-image offense — generally isn't the law that would apply, since its core requirement is a fabricated sexual image of another real person. The picture changes if content resembles a specific real, identifiable individual, which is squarely the scenario Taiwanese law was written to address. None of this is a substitute for legal advice about your specific situation — Taiwan's AI-related law is still developing through 2026, and only a lawyer licensed in Taiwan can tell you how it applies to what you're actually doing.

FAQ

Is there a specific Taiwan law about AI-generated content?

Not one single law covering AI content broadly, as of 2026. Taiwan's most directly relevant statute, Criminal Code Article 319-4, targets a narrower harm — fabricated sexual images of a real identifiable person made without consent — rather than AI-generated content as a general category. This is a general summary, not legal advice.

Is it legal to generate AI images of a real person in Taiwan?

Generating a fabricated sexual image of a real, identifiable person without their consent is exactly what Taiwan's Criminal Code Article 319-4 was written to criminalize, with penalties of up to five years (seven if done for profit). Fictional, non-real-person content sits in a different category, but for anything involving a real person's likeness you should consult a lawyer licensed in Taiwan rather than rely on a general summary.

Does Taiwan require AI-generated content to be labeled?

Not as a general rule for all AI content. Taiwan does require online ad platforms to disclose AI-generated or deepfake material in advertisements under the Fraud Crime Hazard Prevention Act, and its new AI Basic Act (in force since January 2026) sets broader principles that could inform future labeling rules, but there's no universal AI-content labeling law in force yet.

Where can I find the actual text of Taiwan's AI-related laws?

Taiwan's Ministry of Justice publishes the authoritative text of the Criminal Code, including Article 319-4, through the official 全國法規資料庫 (Laws & Regulations Database) — that's a more reliable primary source than any secondhand summary, including this one.