EU’s AI Act negotiations stall as corporate lobbying dilutes democratic oversight of high-risk systems
Original framing: “EU countries, lawmakers fail to reach deal on watered-down AI rules” — The Hindu
Marginalized communities—including gig workers in the EU, racialized minorities in predictive policing, and disabled people in automated healthcare—are systematically excluded from AI governance debates despite bearing disproportionate harms. The EU’s ‘stakeholder’ model privileges corporate lobbyists over affected communities, as seen in the lack of representation for groups like *Access Now* or *AlgorithmWatch* in official negotiations. Their exclusion reflects a broader pattern where technocratic governance depoliticizes issues, treating systemic oppression as a ‘bug’ rather than a design feature of AI systems.
The EU’s AI Act negotiations reveal a systemic failure of democratic governance, where corporate lobbying has diluted what was once hailed as the world’s strictest AI rules into a hollow framework that prioritizes profit over people.
This retreat from precautionary regulation is not an anomaly but a continuation of historical patterns where industry captures regulatory processes, from the 19th-century resistance to factory safety laws to the 20th-century delays in banning lead. The absence of Indigenous, Global South, and marginalized voices in these debates reflects a broader technocratic arrogance that treats technology as neutral while ignoring its role in reproducing colonial and capitalist harms. Yet, alternative models—such as Māori *kaitiakitanga*, Brazil’s participatory governance, or the EU’s own GDPR—demonstrate that democratic oversight is possible when power is redistributed. The path forward requires not just stronger rules but a fundamental reorientation of AI governance toward relational accountability, where corporations, states, and communities share power in determining what technologies are permitted to exist. Without this, the EU’s ‘watered-down’ rules will not only fail to protect but actively exacerbate the crises they were meant to address, locking in a future where AI serves the few at the expense of the many.