As our readers know, this is no longer a crime in the UK, although one rather suspects that other branches of Contempt of Court could find themselves extended to cover some of that ground if – for example – it became routine to denounce Judges as ‘Enemies of the People’ and so on.
It seems that judicial objections are behind a ban apparently being imposed in France on computer analysis of individual Judges’ decisions. Our original source referred to an online journal called the Artificial Lawyer (link to their story here), which leads to the actual text of the law (s.33 Justice Reform Act 2019, to give it a UK-style name – link here) as well as a translation. So far as we can tell, the articles seem to summarise soundly what the Act says – but disclaimers definitely apply. A lawyer with dual US/ French qualifications discusses the matter on his blog here.
We are not sure whether this would ever (as was suggested) have counted as Scandalising the Judges in England & Wales – although since it used to be regarded as a Breach of Privilege and/or Contempt of Parliament to report Parliamentary proceedings, we should perhaps not rule that out entirely. Naturally, much of the comment is along the lines of ‘Which tech companies are going to pay any attention to this?’ – we have not yet seen an analogy with King Canute and the waves, hopefully due to improving popular understanding of the incident rather than the opposite, but that is somewhat the flavour. We share the concern – if that is what is at the root of this – about a ‘Computer Says ‘No” culture where government is by unaccountable algorithms: if the result is ‘wrong’ then that is your fault because ‘Computer says …’
We throw in some comments on this story of our own.
Firstly, and some online comments may have hinted at this, one wonders whether that is the true objection – that data companies will measure model and try to predict Judges’ decisions. Might the French Judiciary be (also) concerned at what the French equivalent of the Ministry of Justice could do with this kind of information? And would they do it well?
Secondly, we are curious how software from common law countries would cope with a civil law system. A late and much-lamented colleague learned in both suggested to us once that what in a common law system counts as law is in a civil law system often considered to be fact. What we understood him to mean was that a civil law court might not consider itself to be applying inconsistent law when it seemed to a common lawyer to be doing so: rather, it was addressing different sets of facts with rightly different results (or vice versa). (Perhaps we put that badly – try some hypotheticals).
We suppose that the ‘artificial intelligence’/ machine learning promoters will say that the latter concern makes no difference, and that the software will work its way around that. Perhaps. Maybe that is so because the software will not (we think) attempt directly to emulate a human jurist’s mind but rather survey all the data in order to observe correlations and what does and what does not seem significant – so although it might not ‘think’ ‘law vs fact’ it might come at the distinction (slippery enough to humans) indirectly and so well enough to work in practice.
Thirdly, this seems like the Legal Realists’ dream. One supposes that they might demand that the software be open source and the data made public, however. How far are we from, ‘M’Lud, we’re grateful. There are instructions to apply for leave to appeal – the data says that you are overruled more often than any other High Court Judge in this kind of case’?