Human Rights and AI: Is Legal Regulation Optional?
The swift development of artificial intelligence (AI) raises pressing legal and ethical challenges for governments globally. Igor Rogov, chair of Kazakhstan’s human rights commission and president of the Kazakhstan Criminological Association, highlights the urgent need for legal frameworks and international cooperation to safeguard human rights in the digital era.
Global Divide on AI Regulation
Countries are split into two main groups regarding AI regulation. The first group, including China and the United States, resists legal restrictions, arguing that regulation could slow scientific progress and reduce national competitiveness. Rogov points out a notable shift in the US stance: "A few years ago, Elon Musk warned about AI's risks to humanity. Today, both Musk and the Trump administration advocate removing restrictions to maintain competitive advantage."
The second group, mainly continental European nations and the UK, supports strong legal oversight for AI development and application. They favor comprehensive legislation addressing responsibility and human rights concerns. Rogov emphasizes that legal frameworks for AI are nearly nonexistent now and insists that legislative restrictions are essential.
Emerging International Frameworks
Several international bodies have introduced ethical and legal standards for AI. The European Parliament passed the Artificial Intelligence Act in March 2024, followed by the Council of Europe’s Framework Convention on Artificial Intelligence, Human Rights, Democracy, and the Rule of Law in May. The United Nations adopted the Global Digital Compact in September, outlining digitalization principles including AI use.
Although EU laws do not directly apply to Kazakhstan, the Framework Convention is open to signatures from non-member states. Rogov notes that Kazakhstan has yet to discuss joining this convention.
Kazakhstan’s Legislative Initiatives
Kazakhstan is advancing its legal framework with a Digital Code and a draft AI law submitted to parliament. Some experts question the need for two separate documents, suggesting AI provisions belong within the Digital Code. For now, the AI law is prioritized, with the Digital Code expected later.
The Human Rights Commission is hosting public discussions on the draft via Telegram. Skepticism exists about the law’s necessity and constitutionality, with some favoring sub-legislative regulation instead.
Three Legal Priorities for AI Governance
- Liability: Determining responsibility for AI-caused harm remains unclear. Should it be the operator, developer, or programmer? Some foreign debates even consider granting AI itself legal personality, but Rogov dismisses this as science fiction at present.
- Intellectual Property Rights: Ownership of AI-generated outputs like research, art, and music is unsettled and inadequately addressed in the draft legislation.
- Combating Fraud and Criminal Use: AI-facilitated crimes, such as voice and facial mimicry for fraud, are increasing. Rogov stresses this as a top legislative priority.
While the proposed law mainly tackles technological aspects, it includes mechanisms for protecting human rights and public safety. Rogov applauds lawmakers for proactively drafting the law and making it publicly accessible on the parliament’s website.
Learning from Global Best Practices
Rogov urges deeper collaboration with domestic and international legal experts, including those from France, Germany, and Russia. He suggests submitting the draft to the Venice Commission of the Council of Europe for expert review, given Kazakhstan’s history of productive engagement with the body.
He also points to Russia’s recent law targeting criminal AI use as a model Kazakhstan should consider, with criminologists and legal scholars contributing to mechanisms that prevent AI-enabled digital crimes.
AI in the Judiciary: A Cautious Start
Kazakhstan’s Supreme Court has introduced AI tools to draft civil case decisions. Parties can review AI-generated drafts before hearings, helping reduce errors and corruption risks. Judges retain full decision-making authority but must justify rulings when AI drafts are involved, increasing transparency.
Rogov notes this approach narrows corruption opportunities and lowers professional errors, suggesting future reforms are needed in criminal, civil, and other legal branches to integrate AI effectively.
Balancing Rights, Innovation, and Sovereignty
Writing effective AI laws is challenging. Rogov points out that while legal theorists can outline goals, drafting practical, enforceable norms is much harder. Laws must protect society, the state, and individuals while supporting technological innovation.
He stresses balancing these objectives is key and calls for harmonized digital legislation within the Eurasian Economic Union (EAEU), respecting each country’s sovereignty. Kazakhstan’s digital policy focuses on people’s needs and interests, aiming to ensure digitalization benefits citizens without restricting their rights and freedoms.