Ukraine's Ministry of Digital Transformation has unveiled a regulatory roadmap for artificial intelligence (AI), with the publication of this roadmap occurring on October 7. The primary objectives of this roadmap, as stated on the ministry's website, are to assist local businesses in preparing for the adoption of AI-related laws that are akin to the EU Artificial Intelligence Bill and to educate citizens on safeguarding themselves against AI-related risks.
The roadmap operates on a bottom-up approach, featuring recommendations that begin with less stringent measures and gradually progress to more comprehensive ones. Its focus is on equipping businesses with the necessary tools to adapt to future regulatory requirements even before any legislation is enacted. To facilitate this, the roadmap initially sets a timeframe of two to three years, during which companies can make the necessary adjustments in preparation for potential legal changes.
Oleksandr Borniakov, the Deputy Minister of Digital Transformation, emphasized the plan to establish a corporate culture of self-regulation. This involves encouraging companies to adhere to a voluntary code of conduct that certifies their ethical use of AI. Another key element is a white paper that will provide companies with insights into the methods, timing, and phases of regulatory implementation.
The roadmap also outlines that Ukrainian draft legislation related to artificial intelligence is anticipated to be introduced in 2024. However, it's worth noting that the EU Artificial Intelligence Act permits national laws to take effect earlier. In June, the European Parliament passed the EU Artificial Intelligence Bill, which, upon implementation, will prohibit certain AI services and products while imposing restrictions on others. Banned technologies include biometric surveillance, social scoring systems, predictive policing, emotion recognition, and non-targeted facial recognition systems. Generative AI models, such as OpenAI's ChatGPT and Google's Bard, will be permitted to operate as long as their output is clearly labeled as AI-generated.
