What is shifting
According to Produktion Online, the EU has extended certain deadlines for high-risk AI with the „AI Omnibus“, an amendment package to the AI Act. According to the article, affected are, among others, AI systems used in areas such as employment, education, critical infrastructure, or access to essential services. The article does not name specific new dates.
No all-clear
According to the article, written by Nick Payne of Confluent, companies gain more time to implement the requirements in practice. However, for some requirements it is still being specified how they should be implemented and how compliance should be demonstrated. The extra time should therefore not encourage waiting.
What the AI Act requires
According to the article, the AI Act sets requirements for high-risk AI including data quality, logging, technical documentation, human oversight, robustness, cybersecurity, and accuracy.
Assessment
The article argues that companies should first record where AI influences decisions, including in software from external vendors. They should also be able to trace, for important decisions, which system was involved and which data was available. If companies address these points early, the extended deadline could ease implementation. Whether this applies to a given organisation depends on how strongly its systems count as high-risk AI.
Sources
Editorially owned by Ideal Syka. Sources and method: Newsroom & method. Tips and corrections: ai@i6eal.de.
