Artificial intelligence is a useful tool for regulatory production, but never a substitute for human judgment. The guidelines for deputies, which were examined on July 30 by the Oversight Committee on Documentation Activities of Montecitorio, chaired by the Vice President of the Chamber Anna Ascani (PD), aim for “a conscious and responsible use of AI.” Final adoption is expected in September.
Read more Cammini, Francesco’s unique itinerary to compete with Santiago de Compostela
Ask the Sun
The questions are automatically suggested by 24Ore AI
based on the content viewed.
AI in Parliament is not a simple tool
This is a 33-page document aimed at protecting organizational and IT security, previewed in Il Sole 24 Ore on newsstands and developed in line with the European and national regulatory framework and the key principles of the legislator’s work. “In the parliamentary context – reads the introduction – artificial intelligence cannot be considered a mere productivity tool. Its use is indeed part of a context characterized by constitutional responsibilities, public relevance of acts, guarantee of pluralism, correctness of information, and protection of fundamental rights.”
Four rules of conduct
A premise is mandatory: the recommendations do not replace security policies or the discipline on personal data protection established by the competent bodies of the Chamber of Deputies; instead, they provide general criteria and rules of conduct. Four fundamental ones are put in black and white. The first: always maintain human control (human in the loop), identifying AI as a support tool and not a replacement for human judgment. Every generated, summarized, or reworked content must therefore be verified before use, especially when intended to be included in public or official acts. Responsibility for the content remains in any case with the person who orders its use.
Read more Tourism, the Tropea elevator becomes «urgent»: 2.5 million allocated
Attention to sensitive data
The second suggestion is to pay attention to what is shared. Personal data, confidential documents, drafts of acts not yet presented, or other sensitive information must not be entered into public AI tools or those not provided by the administration. The data entered could indeed be stored, reused for model training, or made accessible to third parties.
Source verification
The third seems obvious, but it is not. Deputies are advised to always verify sources because AI systems can produce plausible but inaccurate information or information lacking adequate source verification: these are the so-called “hallucinations.” Relevant information must therefore always be checked, and it must be ensured that the generated content does not violate intellectual property rights or other rights.
Read more Thousands of mopeds sold as electric bikes seized throughout Italy, an investigation underway …
