District5524
Joined 3 years ago
Joined 3 years ago
The economy now is very different, no gold or silver standard, less isolation and more globalised product chains and processes, and 7-times as many people.
There is a good chance that a large enough number of people will suffer from this transformation, and they will have to fight with whatever they have, because if they don't, they will risk being left on the roadside and not getting any raise.
People having some direction over how jobs are automated should keep that in mind, otherwise it will be too late to convince the masses of people fearing automation that it is their fault not seeing the positive angles of automation.
The legal systems of common law countries are quite similar in both how they do legal training and in many areas that are important for business. Like in contract law or tort. That means not only similar concepts and foundations (things first taught to law students), but also a similar approach to what is seen as a legal problem, and what kind of answers people expect from a lawyer. Also, what they understand under "legal research" - what do you actually research and what is expected from a qualified lawyer to know by heart, how are you expected to find the legal default in a detailed question, how certain that default is, what is up to professional judgment. But even the differences within common law systems is huge in other areas that go beyond these basics. Like those affected by statutes, codes (like CFR or NYCRR) or even local regulations etc. The popular legal benchmarks do not dare going into these more detailed legal work directions. And these laws change frequently, and they are very different from one jurisdiction to another, even within the US, let alone in other countries building on English legal traditions. (Nobody I know of is building a benchmark that affects these areas of law.) And the diversity is even bigger within legal systems lumped together as "civil law systems". Even those based on Roman law, or on German jurisprudence in the 19th century... There is not much common in how French and German students study law, even if both are members of the EU and many areas of law are harmonised or subject to the same legal acts of the EU and the same EU Court of Justice.
If your LLM has access to up to date legal databases of those countries, commentaries etc., the answers are still surprisingly useful in research. That said, no professional should rely on those LLMs that merely use public legal databases, scraping latest laws and cases via self-operated MCP servers.
I think the main difference is how lawyers from a civil law system use the LLMs compared to someone from a common law system - and that is based on the different training they receive as law students, how they see "law" etc. Maybe there is no such a thing as absolute accuracy in the civil statutes either, but at least civil law lawyers have a stronger notion of "law is knowable". There is a higher likelihood of having one "correct" answer, even when talking about more lifelike, detailed questions. (I'm qualified in both a civil and a common law system but not specifically in German law.)
Due to acting like an irrational gambling machine, I agree it can have unwanted indirect discrimination effect in general. But it will probably not differentiate "on the grounds of religion or belief, disability, age or sexual orientation". It is possible, but that would take a lot of work for the lawyers to prove to the court.
I believe the more interesting part is that the EU AI Act (still not in force in this regard until 2 December 2027). This will be clearly a high-risk AI system: "AI systems intended to be used for the recruitment or selection of natural persons, in particular to place targeted job advertisements, to analyse and filter job applications, and to evaluate candidates".
Which does not mean prohibited, but it could later turn out that LLMs will be excluded from being used in high-risk AI use cases (falling under article 6 with no exemptions).
Considering that none of the standards are published yet, I have absolultely no idea how they will ensure compliance with the following parts of Article 10 when using LLMs for such tasks: "(f) examination in view of possible biases that are likely to affect the health and safety of persons, have a negative impact on fundamental rights or lead to discrimination prohibited under Union law, especially where data outputs influence inputs for future operations; (g) appropriate measures to detect, prevent and mitigate possible biases identified according to point (f)"
I don't think that's technically possible to do so with LLMs in general at the moment, even with the full cooperation of the model providers. Maybe you can do some meaningful audits for smaller models. But the EU AI Act may end up excluding all the generic "using-LLM-but-not-entirely-sure-why" vibe coded approaches from high-risk use cases (in Annex III). Which would make sense.
And for non-common law jurisdictions, even the UK level of publicity of judgments is rarely available - anonymisation is a prerequisite: https://homoki.net/en/2024/02/01/On_Anonymisation_of_court_d...
So, for many countries, it's not just that sealed documents are not accessible, many people in traditional democracies couldn't imagine and bear this level of publicity for every court document like you have in the US.
In Germany, you also have the Oberlandesgericht and the Bundesgerichtshof as a higher venue for civil cases (and the Constitutional Court as well). So, this is not like "the law has changed" at all. (I'm not a German qualified lawyer, but qualified in a close enough continental system, and on my way to be qualified in a common law jurisdiction soon, that's why I dare to comment on this aspect)