I am the author of this essay (a former associate professor at the University of Bucharest, now an independent researcher), and I have worked on an EU legislative proposal regarding this subject: the prevention of the capture of one’s personal (or institutional) 'thought line'

An AI model is a reproducible computational model. Everything it performs in the world can be channeled through a closed list of operations—written by humans—effectively acting as a door. What remains outside this gateway is the record of what humans tell the machine (and how...); this record is the product: the capture of the thought line (ThoughtLine)

Why now? Within the EU, the GDPR is being amended/discussed via the "Digital Omnibus" package, and this proposal has been registered with the European Parliament as Petition No. 2969/2026.

An open question: at what point do you think the concept of a 'the closed-list idea' breaks down—during the execution of the code, or through the use of external tools?