Legal regulation of artificial intelligence

Legal authorities and regulators in Taiwan and the United Kingdom have issued warnings regarding the misuse of artificial intelligence in legal practice.

In Taiwan, the Taoyuan District Court dismissed a civil suit after a non-lawyer plaintiff submitted a filing containing a draft generated by consumer-grade AI. The presiding judge explicitly instructed litigants to “do not ask AI, do not use AI to draft pleadings.” Legal experts noted that while AI is becoming foundational infrastructure in other jurisdictions, technical risks such as “hallucinations” and critical omissions remain significant concerns for laypersons.

Similarly, the Solicitors Regulation Authority (SRA) in the United Kingdom warned law firms about the risks of inaccurate legal information and the exposure of confidential client data. The SRA reported receiving 42 reports of potential AI misuse between July 2025 and July 2026, including instances where generative AI produced fictitious case names and citations. The regulator emphasized that AI use does not relieve solicitors of their professional obligations or responsibility for the accuracy of their work.

In the United States, the rise of generative AI has been linked to an increase in pro se (self-represented) federal court filings, particularly in simple civil cases. However, research indicates that this surge in filings has not resulted in higher success rates for these litigants. Furthermore, the emergence of new liabilities is driving the development of specialized AI insurance to address risks such as algorithmic bias and intellectual property exposure.