Verso un legislatore non umano? Brevi riflessioni su alcuni problemi di diritto computazionale

di Stefano Pietropaoli

SOMMARIO: 1. Considerazioni introduttive. - 2. Un (falso) problema: la personalità giuridicamente rilevante degli enti dotati di AI. - 3. Il (vero) problema: il diritto è computabile? - 4. Avvocati, magistrati e notai: professioni computabili? - 5. Dal legislatore umano al legislatore non umano (o transumano)?

The first IT applications to law concerned merely auxiliary functions. But over time, the use of computers in the legal field has become susceptible to increasingly sophisticated functions. It is no longer a question of mere support tools for the legislative function, as could be software solutions dedicated to drafting or AI applications to assess the impact of new provisions introduced in a given legal system. Artificial intelligence could be used for purposes other than the expression, dissemination and amendment of human legislation. A possible scenario is one in which the formation of the law is entrusted ab origine to intelligent systems: the problem of translation from natural language to machine language would thus be definitively overcome, because the law would be born as a code immediately comprehensible to the machine. It would therefore be a matter of rules produced by machines for machines. But would that normative framework still be ‘law’?