AI could break legal systems by finding previously hidden loopholes

What happened

Felix Ringe, a visiting fellow at Stanford Law School, and his colleague Linus Folkerts wrote a blog post. They say large language models (LLMs), the AI systems behind chatbots, are now good enough to find weak spots in written law. They call these weak spots loopholes. A loophole is a gap in the rules that someone can use in a way the lawmakers did not plan.

Key facts

• The blog post was published the week before the article, which is dated 2 November 2026.
• The authors shared several possible gaps in German law that they found with LLM help.
• They only shared gaps they believe could not cause serious harm.
• One example involves an amendment to Germany's Occupational Diseases Regulation (BKV). The authors say it leaves out a required legal reference.
• Another example involves a transit visa rule. It may be invalid because the government did not name the law that allowed it.

The idea behind it

Computer code can have bugs, and hackers can use them. Lawrence Lessig wrote about this idea in 1999, saying code is a kind of law in cyberspace. Ringe turns the idea around. He says law is a kind of code. Both have strict wording and rules. But courts are not like computers. Court results can change from case to case, which makes the comparison imperfect.

Why it matters

If AI can find many gaps quickly, courts may not be able to fix them all. The authors warn that lawmakers may need to make the law-making process much stronger. They suggest that lawmakers could even use advanced AI themselves to check new laws before they pass.

What to watch

• Whether courts in Germany or other countries test the gaps that the authors describe.
• How lawmakers change the way they write and check new laws.
• How lawyers use AI tools in daily work, as one California employment lawyer notes that AI use in law is growing fast.