The Commission on Monday launched an AI whistleblower tool, meant to give insiders at AI model developers a confidential way to share information when they believe their employers are violating the EU’s AI rules.
However EU legal protections for users of the tool will only apply from August next year – potentially leaving those reporting AI concerns at the mercy of their employers should their disclosures be found out in the meanwhile.
While the EU has a whistleblowing law, the part of the AI Act that states this existing law applies to issues in scope of the AI rulebook will only come into force in the second half of 2026.
Despite the disjointed sequencing, Karl Koch, founder of the AI whistleblower initiative – a group advocating for stronger rules to protect people reporting concerns about AI – described the new EU tool as “an incredibly valuable step”.
“It’s the first channel of its kind globally and it looks like they’re taking this quite seriously,” Koch told Euractiv, while noting that the Commission consulted his group for expertise as it developed the tool.
The EU’s whistleblower law already covers product safety in general, he said, implying that many potential AI problems could fall under this wider provision. Since the AI Act is a product safety law, he also suggested that a capable lawyer could argue it falls under the general protections.
Koch also said that it’s not completely unusual for a whistleblowing channel to accepts tips from people who the channel’...
Lawyer Kang Yong-seok, a former operator of the YouTube channel ‘Hoverlab,’ has belatedly been revealed to have apologized for past activities, including raising allegations of sexual assault agai...