3 ms·

A friend of mine is regularly mentioning chatGPT as is lawyer. Given many countries share the same language, I am struggling to understand how he make sure cha

A friend of mine is regularly mentioning chatGPT as is lawyer. Given many countries share the same language, I am struggling to understand how he make sure chatGPT is not hallucinating and that he is basing his advice on the laws of the correct country.

iinnPP 1y ago
It will direct you to the law and language if you ask. A problem not so easily solved here is the difference between the law and reality. Where the LLM is completely blind to which laws are actually enforced in a proper manner.
mrtksn 1y ago
Trust but verify. It will come up with ideas and claims, google all that for authoritative sources and people who went through all this. Ask again for anything contradicting you can find and ask for edge cases you can think of. Change the wording and ask again to see if it gives you the same advice. IMHO hallucinations are not that bad when you have a human in the loop.
- benchly 1y ago
  I think the term is "confabulations" now, which makes a bit more sense to me, but regardless of what we call the hallucinations, it starts with the human knowing how to navigate the process, having at least some basic understanding of the system within which they're using the GenAI (in this case, law/legal) and the skills to cross-reference and verify without using the GenAI itself. If we start using the GenAI to both get the answer and verify it, we're basically creating a feedback loop that probably isn't always going to stand up to scrutiny, despite the fact that GenAI has gotten quite good at these sorts of things. We're not there, yet, but we will be, probably within our lifetime.
- sebastiennight 1y ago
  > google all that for authoritative sources With Google progressively replacing actual websites with AI-generated results, you might come full circle to having Gemini validate the same hallucinations the Gemini app produced in the first place.

iinnPP 1y ago

It will direct you to the law and language if you ask. A problem not so easily solved here is the difference between the law and reality. Where the LLM is completely blind to which laws are actually enforced in a proper manner.

mrtksn 1y ago

Trust but verify. It will come up with ideas and claims, google all that for authoritative sources and people who went through all this. Ask again for anything contradicting you can find and ask for edge cases you can think of. Change the wording and ask again to see if it gives you the same advice. IMHO hallucinations are not that bad when you have a human in the loop.

benchly 1y ago
I think the term is "confabulations" now, which makes a bit more sense to me, but regardless of what we call the hallucinations, it starts with the human knowing how to navigate the process, having at least some basic understanding of the system within which they're using the GenAI (in this case, law/legal) and the skills to cross-reference and verify without using the GenAI itself. If we start using the GenAI to both get the answer and verify it, we're basically creating a feedback loop that probably isn't always going to stand up to scrutiny, despite the fact that GenAI has gotten quite good at these sorts of things. We're not there, yet, but we will be, probably within our lifetime.
sebastiennight 1y ago
> google all that for authoritative sources With Google progressively replacing actual websites with AI-generated results, you might come full circle to having Gemini validate the same hallucinations the Gemini app produced in the first place.