Photo from the internet. The Cabinet has established new standards for the safety and quality of AI systems
According to the resolution, the following key aspects have been approved:
- methodology for assessing the safety of AI systems;
- rules for risk management for high-risk systems;
- criteria for transparency, explainability, controllability, accuracy, reliability, and digital resilience;
- data quality requirements for high-risk AI;
- standards for technical documentation required by developers and owners of such systems.
The document obliges all owners of AI systems, regardless of ownership form and departmental affiliation, to adhere to the following requirements:
Comply with the measures outlined in the resolution.
Conduct safety assessments of AI at all stages of their lifecycle.
In the event that the system is deemed high-risk:
- ensure full compliance with mandatory requirements;
- implement a risk management system;
- develop and publish technical documentation, as well as ensure access to related documents in accordance with the Digital Code.
This resolution will come into effect 15 days after the ratification of the Digital Code.
