Meta’s Muse Spark Consent Design Undermines Enterprise Data Governance and Practical Control Over AI Training Data
Meta places consent for using user uploads as AI training data in a way that common enterprise security and monitoring tools cannot easily detect. If consent mechanisms are pushed into obscure product tiers and screens that evade standard policy and DLP checks, enterprises lose practical control over what content becomes model training fuel, and regulators’ reliance on ‘choice’ becomes largely performative...

