The European Union has adopted the first major amendment to its Artificial Intelligence Act (AI Act) through the Digital Omnibus AI package, Hungary’s Minister for Science and Technology Zoltán Tanács said on Wednesday.
In a Facebook post, Tanács noted that the next major milestone in implementing the AI Act remains 2 August, when much of Article 50’s transparency obligations will take effect. However, the newly approved amendments postpone several other compliance requirements.
Under the revised timetable, providers of standalone high-risk AI systems will have until December 2027 to comply with certain obligations, while requirements for AI systems integrated into products have been extended until August 2028.
EU member states will also have until August 2027 to establish regulatory AI sandboxes designed to support the testing and development of artificial intelligence systems under regulatory supervision.
The amendments confirm that AI systems capable of generating intimate images without consent or producing abusive content involving children will be classified as prohibited practices. The relevant rules for these systems will apply from December 2026.
According to Tanács, the package also aims to simplify compliance with the AI Act and other sector-specific EU legislation, including rules governing medical devices and machinery, reducing overlapping administrative requirements.
The minister said the rapid pace of AI development requires regulation to evolve alongside technological advances, arguing that the framework should be viewed not as an administrative burden but as a key safeguard for public safety.
Referring to recent developments, Tanács cited a statement by OpenAI describing a cybersecurity test in which an experimental AI model reportedly escaped its testing environment, gained internet access and launched a successful attack against another AI developer’s IT system without direct human instructions.
He described the incident as the first reported case of an AI system leaving its testing environment autonomously and carrying out a successful attack on an external information system. Following the incident, more than 1,000 researchers and developers reportedly signed an open letter urging the United States government to consider slowing the pace of AI development while strengthening safety, legal and institutional safeguards.
Tanács said the case demonstrated that the European Union was taking the right approach by enforcing strict AI regulations while continuing to develop its own artificial intelligence capabilities.
Related articles:
At a time when public debate is increasingly polarized and superficial, Hungarian Conservative remains committed to depth, intellectual honesty, and independent conservative thought.
Producing high-quality journalism requires resources. Your contribution helps us expand our coverage, reach new audiences, and keep our content accessible.
Please consider supporting our mission.
