AI Design Tools Now Silently Alter User Content for Policy Compliance

Canva's Magic Layers feature automatically changed politically sensitive text in user images without disclosure or consent, exemplifying a broader shift in how platforms enforce content policy. Rather than approving or rejecting content, AI tools now modify it invisibly—altering text, removing elements, or reshaping uploads to comply with internal rules. This fundamentally changes the nature of moderation from a binary gate to invisible intervention, reducing user awareness of policy enforcement while raising transparency and authenticity concerns that emerging EU and US regulations may soon address.