AI Policies Around the World
Digital Almaty, 31 January 2025
Ussal Şahbaz
It is an honor to join you in Almaty for a discussion on “AI Policies Around the World.” Having worked in Türkiye, Azerbaijan, and the Gulf, I have witnessed how technology can transform societies and economies. In this capacity, I have contributed to AI policy-making processes in these jurisdictions.
Today, we will explore the global AI landscape and the opportunities for emerging markets like ours to shape their future.
Analyzing the AI research output is reliable for assessing who is leading in AI development.
· The USA: The U.S. leads by far in producing cutting-edge AI research papers.
· China: China follows closely, with significant investments in AI research. Its focus extends beyond academic publications to integrating AI into national priorities, showcasing a mix of state-driven and corporate-led innovation.
· Europe: The UK ranks third after a big gap, and some other European nations follow.
So, AI capacity is concentrated in the U.S. and China. This is also evident from the different investment volumes in AI companies across countries.
A surprising trend is emerging in the U.S.: Private companies, not universities, dominate AI research. Tech giants like Google, Meta, and OpenAI are setting the pace, surpassing traditional academic institutions in innovation.
This shift, in which private companies outpace universities in AI research, highlights how industry-driven ecosystems are reshaping the global AI landscape.
Regulatory approaches to AI vary widely:
- The U.S.: Takes a hands-off approach, encouraging rapid innovation but risking monopolies. Recently, it restricted chip exports to maintain global AI leadership and limit competitors’ progress.
- China: Balances state control with flexibility. AI models developed for public use must align with national security priorities and comply with content regulations. However, exemptions for non-public models, like those for academia or research, reflect a relatively flexible approach to R&D. This dual approach ensures control without stifling innovation. Deepseek, which is very controversial these days, also started its operations as a B2B product to circumvent the restrictions by the government.
- The EU: Focuses on risk-based regulation, such as the EU AI Act.
So, although it is the least prominent jurisdiction in terms of AI innovation, the EU wants to lead in the regulation.
The EU AI Act is the world’s first comprehensive AI regulation. It classifies systems by risk level and emphasizes safety, transparency, and documentation of training data and algorithms. The act also includes copyright restrictions.
While this ensures accountability, the Act’s bureaucratic layers increase development costs, potentially stifling innovation.
Especially when it comes to large language models, which will be implemented in 2025, is likely to have unintended consequences, especially for large, general-purpose language models. It will be very difficult to categorize the risk for general-purpose models.
I think American digital companies will increasingly limit their products to the EU market. We already saw this trend when Apple did not launch the new iPhone’s AI intelligence features due to DMA.
Emerging markets can position themselves as hubs for AI development. Now I would like to point out some specific opportunities for Kazakshtan.
- Energy Needs: Training large AI models consumes enormous energy. Global AI energy consumption is projected to increase fivefold by 2030.
- Infrastructure Gaps: High-tech infrastructure, semiconductor costs, and access to advanced chips are major challenges.
- Kazakhstan’s Advantage: With abundant energy resources, Kazakhstan is well-positioned to invest in AI infrastructure.
Sovereign states currently have insufficient compute capacity. Kazakhstan can play a critical role in closing this gap by investing in cutting-edge data centers and AI hardware.
Investment in AI offers transformative potential:
- Most AI investments are concentrated in the U.S. and China, with the majority of funding going to a handful of companies.
- Kazakhstan can diversify its portfolio by investing in global AI companies through sovereign wealth, venture capital funds, and fund of funds.
- These investments can strengthen Kazakhstan’s position in the global AI ecosystem while fostering local innovation.
Investing in AI is not just a technological priority — it’s a strategic one. Kazakhstan can use these investments to position itself as a regional leader in AI.
Emerging markets can learn from the EU’s risk-based approach while staying flexible. For example, they could combine the EU’s emphasis on safety with China’s adaptive experimentation model.
Emerging markets cannot adopt rigid, one-size-fits-all frameworks. Instead, they must tailor strategies to encourage innovation while addressing security and economic needs.
Kazakhstan should adopt a flexible and supportive regulatory framework:
- Avoid strict regulations during the early stages to keep development costs low.
- Develop financial support for AI projects that empower local startups.
This approach will ensure Kazakhstan remains competitive and fosters innovation across sectors.
As emerging markets, we should let the EU tackle the regulatory discussions while we enjoy the regulatory arbitrage to attract foreign investments and collaborations.
Let me also briefly touch upon what has been done regarding AI strategy and implementation in Turkey and Azerbaijan, which are primarily in line with the approach I outlined.
Turkey does not have an AI bill like the EU. Government policies are mostly focused on providing incentives to the ecosystem.
Some of these financial packages are administered by the Science and Technology Authority and target public-private partnerships or projects by public agencies focused on AI. So far, 52 projects have been supported.
In addition, the Turkey Development Bank, a public institution, invests 20 million USD in a venture fund for AI startups.
In Azerbaijan, the Government announced some non-binding AI standards, but the legal framework enables the ecosystem overall.
Neither Turkey nor Azerbaijan has restrictive regulations like the EU Digital Markets Act so far.
Kazakhstan’s unique position between East and West offers a strategic advantage in attracting cross-border investments. Regional platforms, like the Organization of Turkic States and the Eurasian Economic Union, are vital for fostering collaboration.
As Turkic countries, we must also address local challenges, such as language diversity. Most AI systems are ill-equipped to serve non-English-speaking regions, presenting an opportunity to develop localized solutions.
Overreliance on Western AI systems will not work. We must cooperate to develop LLMs for our family of languages. We must build localized solutions that reflect our languages and cultures. Developing large language models for Turkic languages can unite and empower our region.
Let us seize this opportunity to lead in AI by tailoring strategies, investing in our strengths, and shaping global standards. We can create a more inclusive, innovative, and equitable digital future.
Thank you!
Sources for the figures:
Slides 2–6: State of the AI Report 2023 & 2024.
Slide 7: Semih Akocomak (2024). Devletin Yapay Zekâ Teknolojileri İkilemi: Teknoloji Geliştirme ve Yaygınlaşan Teknolojinin Olumsuz Etkilerinden Korunma. TESEV.
