The EU AI Act, effective August 2024, mandates a Code of Practice for AI developers. Still, current drafts weaken human rights protections by making risk assessments optional for many categories, including fundamental rights and discrimination. This change, influenced by corporate interests, jeopardizes rights amid widespread AI use. International standards emphasize risk assessments for human rights, highlighting a disconnect between the Code and global norms. The draft needs revisions to align with robust protections for human rights to maintain the EU's leadership in AI governance.