I consulted one university in British who have similar data privacy laws, and the reply I got from their admission office indicates that if I indicate explicitly in the letter or upon providing it, the letter won't be shared with the student without consulting me first (but it's theoretically possible that the university may decide to share the letter anyways with my disagreement, if they see proper to do so). That is, you will have to explicitly indicate that you at least must be asked before releasing the letter to students, in theory.
Whenever you're not sure, you can simply write an email to the program's admission office and ask for an official answer/promise.
Personally, I am someone who either writes a nice letter or refuses to recommend, but I always insist that students I recommend must waive their rights to access my evaluations.
The arguments that some comments above on how students can benefit from reading the reference letter are unconvincing. If someone, as the letter writer, feels comfortable sharing advises with the student, he/she can just tell the student face-to-face or directly over emails. Recommendation letter is certainly not the only chance that students can learn about their strengths and weaknesses, and if it is one, it's usually far from the best way to communicate (not only because the letter is not written to the student as its target reader, but also the writer and a student reader may have different interpretations, because the professor have his/her own comparison groups so the evaluations may not match the student's expectation, but this could be better understood by the recruiters especially when the professor has been writing letters to this recruiting program in several years already). The form of a confidential letter not only ensures the recruiters that this letter faithfully expresses the opinions of the letter writer, reducing their concerns of impacting the interpersonal relationship to the minimum level, but moreover, (the letter) may contain confidential information (even it's just summary) regarding the performance of other students which cannot shared with the subject of the letter. I usually tell my students roughly what I wrote in the letter, but I never let them see the letter per-se.
The recommendation process needs to be built on some mutual consensus from the professor and the student. If you don't like the term of waiving the access, don't ``sign the contract'' by requesting my recommendation -- I am by no means the only possible person for students to ask.