Legal Prompting - Chain-of-thought e few-shot prompting nel legal

Two techniques change the way a model approaches a legal problem: chain-of-thought — the explicit request to lay out the logical steps before the conclusion — and few-shot prompting — providing two or three well-chosen examples to steer format and method of the answer.

In this episode we look at:

• how to structure step-by-step reasoning with a concrete example on data transfers outside the EU (Chapter V GDPR, Schrems II);
• how to pick few-shot examples without introducing bias;
• how to combine the two techniques for the analysis of complex decisions;
• the three limits to be aware of: context length, example bias, and plausibility that is not legal correctness;
• why documenting prompt, examples and verification is already AI governance.

Chain-of-thought and few-shot are not tricks: they are the way we translate our legal method into instructions understandable to the model.

In the next episode we will apply these techniques to the analysis of contracts and clauses.

🎧 Also on Apple Podcasts, Spotify, Amazon Music, and Podcast Index.

🌐 nicfab.eu