Large language models may help researchers find gaps in legislation that could be exploited, a Stanford-affiliated legal scholar and a collaborator argue. Their examples from German law turn the familiar idea that code has bugs on its head: what happens when someone starts looking for bugs in the law at scale?
Watch Desk analysis
What happened
The Register reports that visiting fellow Felix Ringe and legal scholar Linus Folkerts used LLMs to identify possible statutory gaps in Germany. Among the examples they describe are an amendment that may omit a constitutionally required reference to its legal authority, and a transit visa rule whose authority may not have been properly identified. They say they are sharing examples they believe would not cause serious harm.
The researchers argue that courts may struggle to resolve a growing volume of adversarial challenges. But the examples are potential gaps, not proof that a court would accept a challenge or that the laws are invalid. Legal experts quoted by The Register say human review is needed to establish whether a gap is real and whether it has meaningful consequences. The Register’s report sets out the examples and the responses.
Why it matters
AI tools are already being discussed as assistants for legal research and drafting. This is a different use: searching legal text for weaknesses that a person might turn into a challenge. If that search becomes cheap and repeatable, lawmakers and courts could face more work separating genuine defects from clever-looking but legally irrelevant arguments.
That is not the same as an AI discovering a magic loophole. Courts interpret statutes, and legislators can amend them. The real question is whether automated searching changes the volume or speed of challenges enough to put pressure on those processes.
Our read
This is a useful warning about a plausible capability, not evidence that legal systems are about to collapse. The German examples make the argument concrete; the experts’ cautions make clear why “the model found a gap” is only the start of the legal analysis. AI may help check legislation, but a confident answer is no substitute for a court deciding what the text means.
What to watch
- Whether the researchers publish further examples or methods that can be scrutinised.
- Whether courts or legislative counsel report using AI to identify drafting gaps.
- Whether any of the proposed German examples lead to formal challenges or authoritative rulings.
Discussion spark: Should lawmakers routinely use AI to look for gaps in draft legislation, or would that risk giving automated analysis too much influence over what counts as a legal flaw?
Sources and evidence
- AI could break legal systems by finding previously hidden loopholes – The Register (2 November 2026, 21:12 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.