August 2 marks the EU AI Act’s Article 50 disclosure deadline, requiring employers to tell workers and candidates when they’re interacting with AI and to label AI-generated content. The heavier employment-specific rules covering hiring tools, performance systems, and workforce monitoring were pushed to December 2027 under the EU’s Digital Omnibus package, but that runway is narrower than it looks. Only 13% of companies globally have any public AI governance framework, per a Thomson Reuters Foundation and UNESCO analysis of nearly 3,000 companies<\/a>, and just 7.4% of firms using AI in HR consult diversity staff on those projects.<\/p>\n Any employer using AI in recruiting, performance reviews, or workforce monitoring inside the EU is already out of compliance on literacy requirements that took effect in February 2025, and the August 2 disclosure rules hit whether or not the vendor warned you. The exposure isn’t evenly distributed. Organizations that bought AI recruiting or monitoring tools from third-party vendors and assumed compliance was the vendor’s problem are the ones most exposed. Under the Act, deployers, meaning the employers themselves, are required to follow vendor instructions for use, which means you have to know what those instructions say.<\/p>\n The 2027 deadline for high-risk employment AI creates a false sense of distance. The definition of “high-risk” under the Act covers nearly every tool already common in HR stacks: resume screeners, candidate scoring, targeted job ad placement, performance monitoring, task allocation systems. That’s not a niche list. It’s the core of modern talent operations. Companies that use the 2027 date as a reason to defer governance work will spend the intervening years accumulating undocumented deployments, vendor dependencies without audit trails, and employees who’ve never been told they’re interacting with AI systems making decisions about their careers. Fixing that retroactively costs more than building it right the first time.<\/p>\n The 2% figure for internal AI complaints channels is the number that should follow CHROs into their next board conversation. Regulators treat the absence of an employee feedback mechanism as evidence of structural non-compliance, not just an operational gap. The CHRO who can point to a functioning complaints channel, documented vendor instruction reviews, and even a basic AI literacy program for HR staff is in a categorically different position than one who can’t, regardless of where the 2027 deadline falls. That’s the budget case for moving before the deadline pressure arrives, not because the Act demands it by August, but because the audit clock is already running.<\/p>\nWhat this means for your business<\/h2>\n