AI Autonomy Prohibition in Kazakhstan: Government Sets Restrictions on Self-Governing Artificial Intelligence
In a move to tackle the looming AI invasion, Kazakhstan's parliament, the Mazhilis, has given its initial nod to a draft law on AI. According to Vechernyaya Astana, the legislation aims to address key security concerns.
This new law divides AI systems into three categories: assistants, semi-autonomous systems, and fully autonomous algorithms. The latter will face a total ban within the Republic of Kazakhstan. Let's break it down:
- Autonomous Algorithms: These bad boys can take care of tasks or make decisions on their own, sans human intervention. They rely on AI and machine learning for self-learning, interacting with the environment, and adjusting their behavior accordingly. Think self-driving cars, drones, and helper robots.
The law also emphasizes security by classifying high-risk AI systems as critical information infrastructure objects. This move enables greater control over their security.
Moreover, the bill outlaws several AI usage practices that step on our toes or play dirty tricks:
- Using AI to manipulate people's behavior without their conscious consent
- Leveraging AI to exploit vulnerabilities such as age, disability, or social status
- Assessing or categorizing people based on their social behavior or attributes (so-called social scoring)
- Analyzing people's emotions without their consent
- Employing AI for real-time facial recognition in public spaces, unless permitted by law
- Creating and spreading AI-generated outcomes prohibited by law
A crucial aspect of the draft is the provision for compensating users if AI systems cause harm. The owners and operators of these systems will be held responsible. Furthermore, the law suggests insurance for AI-caused harm as a voluntary measure, which could become a critical element in safeguarding citizens' rights.
The global trend in AI regulations focuses on transparency, reducing bias, and maintaining safety. Similar to the EU AI Act, Kazakhstan's regulations might adopt a risk-based approach to regulate AI systems. Meanwhile, state-level AI laws in the U.S. often delve into transparency, fairness, and accountability.
Though specifics about Kazakhstan's regulatory approach for AI are still sketchy, ongoing efforts are underway to study and analyze generative AI regulations in Kazakhstan through comparative legal analysis. A thorough understanding can be gained by consulting recent legal documents or official announcements from Kazakhstan's regulatory bodies. Stay tuned for more updates!
- The new AI law in Kazakhstan, aimed at addressing security concerns, classifies autonomous algorithms as critical information infrastructure objects, allowing for greater control over their security.
- In a bid to maintain transparency and reduce bias, the law outlaws several AI usage practices such as manipulating people's behavior, exploiting vulnerabilities, and assessing or categorizing people based on social behavior or attributes without consent.
- As part of the draft law, there is a provision for compensating users if AI systems cause harm, and the law suggests insurance for AI-caused harm as a voluntary measure to safeguard citizens' rights.
