La CNIL rendra-t-elle sa première décision d’application de l’IA à haut risque selon l’AI Act avant le 2 août 2026 ?

Pourquoi : The supplied evidence shows the EU’s specific obligations for high‑risk AI systems were deferred beyond 2 August 2026, with multiple independent sources stating those high‑risk requirements apply from 2 December 2027 (or later). For example, TechGDPR explicitly reports that obligations for high‑risk AI systems were postponed until 2 December 2027; Dentons (publication 2026-07-28) likewise reports the Omnibus postponed compliance deadlines for stand‑alone high‑risk systems to 2027; Pearl Cohen’s summary also records the deferred implementation timeline to 2 December 2027. No page in the provided evidence reports the CNIL issuing any formal enforcement decision taken under the AI Act’s high‑risk systems provisions by 2026‑08‑02. (The CNIL’s high‑profile May 2026 IQVIA sanction appearing in the evidence was under health data / GDPR rules, not the AI Act high‑risk provisions.) Because the high‑risk provisions were not in force by 2026‑08‑02 per multiple corroborating sources, and there is no evidence in the supplied material of any CNIL enforcement decision under those AI Act high‑risk provisions by that date, the correct settlement is NO.