Joint Open Letter on Rejecting the Proposal to Undermine Transparency in the AI Act
In a significant push for accountability within the realm of artificial intelligence, CDT Europe, along with 60 organizations, has signed an open letter initiated by Access Now and European Digital Rights (EDRi). This letter, directed towards the EU institutions, strongly urges the rejection of the proposed deletion of the AI Act’s Article 49(2), which serves as a crucial transparency safeguard for high-risk AI systems.
The Loophole in the AI Act
The AI Act has faced criticism for allowing providers of AI systems categorized as high-risk to opt out of obligations based on broad criteria. This controversial loophole raises alarms about the potential for unregulated AI systems to proliferate. A vital safeguard against this damaging carve-out is the requirement for these providers to register their AI systems in an EU database. This measure enables necessary scrutiny and oversight over potential exploitations of the loophole. However, this transparency safeguard is currently under threat.
The Importance of Transparency
Maintaining transparency in AI system operations is crucial to ensure that providers cannot circumvent the core obligations laid out in the AI Act. The removal of the transparency obligation poses significant risks, allowing potentially dangerous AI systems to proliferate without oversight. As highlighted in earlier briefs regarding the AI Act Omnibus proposal, the failure to disclose an AI developer’s self-assessment leaves regulators and users without the means to make informed decisions. This lack of information can hinder their ability to investigate or avoid potentially harmful AI systems.
The Human Cost of Removing Safeguards
Eliminating the transparency obligation will incur a significant human cost in exchange for trivial cost savings for AI developers. This move threatens to undermine the enforceability of the AI Act drastically and worsen risks related to health, safety, and fundamental rights.
In conclusion, the call to reject the proposal that undermines the transparency safeguards of the AI Act is not just a technical concern; it’s a matter of public safety and accountability. The future of AI must be grounded in principles that prioritize human rights and ethical governance.
Read the full letter here.
