Azernews.Az

Nazrin Abdul

Read more

As the AI Impact Summit unfolds in Delhi bringing together global tech leaders such as Google CEO Sundar Pichai, OpenAI CEO Sam Altman, and Anthropic CEO Dario Amodei, a broader global debate is taking shape. India is positioning AI as a tool for social transformation across the Global South, while U.S. companies continue to emphasize commercial scale, automation, and technological supremacy.

In this context, Azerbaijan is accelerating its own digital transformation. The local government is even supporting the development of more advanced platforms to enable a range of institutions to use AI more effectively. However, leveraging these platforms on the global stage, acquiring experience, and, in short, establishing a position in the country’s rapidly polarising AI landscape, is a crucial matter that is both thought-provoking and of interest to everyone. AzerNEWS spoke with Etibar Aliyev, an AI expert leading large language model (LLM) research and evaluation to better understand the strategic direction Azerbaijan could take.

Rather than framing AI as a race for the largest models, the expert argues that Azerbaijan’s opportunity lies in disciplined execution and measurable deployment. He believes that, based on his current leading LLM research, testing, and benchmarking for models like BERT, Llama, and Gemini, the lesson is clear: the winners aren’t the countries that chase the biggest model - they’re the ones that build the strongest apply–measure–scale engine and turn AI into national productivity.

“For Azerbaijan, the strategic direction I’d pursue is applied AI at national scale with governance as a differentiator. That means investing less in prestige compute and more in the delivery stack: secure data pipelines, reusable RAG components, MLOps, and crucially, independent evaluation. We should have a national benchmark and safety practice so every ministry and major enterprise can answer: Does this model work reliably in Azerbaijani? What’s the hallucination rate? What’s the cost/latency? What are the failure modes?” the expert stressed.

According to him, competing globally is then about specialization: efficient, localized AI - domain-tuned assistants and agentic workflows embedded into real operations. "This is exactly how I’ve seen value created: measurable improvements, then scaled rollout," he said in a comment. “If Azerbaijan couples that with responsible AI governance - model risk management, audits, and clear procurement standards - it can become a regional exporter of applied AI solutions, not just a consumer,” he added.

With Indian Prime Minister Narendra Modi positioning India as an AI hub for South Asia and Africa, Azerbaijan is exploring its potential to become a regional AI hub in the South Caucasus and identify which sectors could drive this leadership.

The expert suggests redefining what a “hub” actually means - not as a branding exercise, but as a place known for repeatable, scalable deployment. E.Aliyev emphasized that to become such a hub, Azerbaijan should focus on building repeatable reference implementations. He noted that once a scalable model - covering data, tooling, evaluation, and change management - is successfully demonstrated, others are likely to follow.

According to him, sector-wise, the strongest “hub builders” are:

• Energy and industrial operations: predictive maintenance, anomaly detection, and safety monitoring.
• Logistics and trade corridors: document automation, customs workflows, routing, and demand forecasting.
• Digital public services: visible improvements in citizen experience that can serve as a regional blueprint.

The AI expert added that the goal should be to position Azerbaijan as the region’s “deployment factory” - with shared tooling, shared evaluation standards, and shared governance frameworks - enabling startups and government institutions to implement AI solutions faster and more safely than neighboring markets.

While AI in India is being applied in education and multilingual accessibility, Azerbaijan is assessing which sectors hold the greatest potential for AI-driven development today.

The expert presumes focusing on areas where Azerbaijan can get near-term wins without waiting for perfect data or frontier research: "First, high-volume service delivery, including public services, telecommunications, and banking, where AI applications such as multilingual assistants, enterprise search, document extraction, summarisation, and agent-based workflows with human oversight can generate relatively quick returns.

Second, education and workforce upskilling, particularly localized solutions. He indicated that an Azerbaijani-language tutoring layer combined with job-skill copilots could deliver measurable impact, provided that implementation is tied to clear performance indicators.

Third, industry and infrastructure, particularly edge and operational AI. He pointed out that use cases such as computer vision inspection and predictive maintenance have already demonstrated practical value in multiple markets," he said.

He also added compliance and risk automation, including KYC/AML processes, sanctions screening, and onboarding, as areas where AI could reduce manual workload, provided that proper governance and oversight mechanisms are in place.

With Azerbaijan recently accelerating digital reforms, including integrating AI into ministries, expanding language support through global AI platforms, and launching Gemini in Azerbaijani, along with advanced creative tools like Nano Banana and Veothe, the country is positioning itself for meaningful AI growth.

“They can, because this is the right combination: language access, digital rails, and institutional ownership,” he said, adding that the rollout of Gemini in Azerbaijani reduces adoption barriers for citizens and civil servants while introducing creative and multimodal tools to the market.

The expert also emphasised that unified digital architecture and ministry-level accountability create the conditions for scale. However, he strongly believes that the decisive factor is execution. To turn this into a functioning ecosystem, he recommended three concrete steps, including a shared government AI platform (secure RAG, model gateway, logging, and monitoring), a national evaluation and safety playbook (benchmarks, red-teaming, and audit trails), and a delivery scoreboard tracking cost savings, reduced time-to-service, error rates, and citizen satisfaction.

“If these three elements are implemented effectively, then yes - these developments can position Azerbaijan to move from initial adoption toward a scalable AI ecosystem,” the expert said.

While the AI Impact Summit in Delhi highlights the growing global competition for technological leadership, Azerbaijan’s path is more likely to differ from that of major AI powers. Rather than competing in frontier model development, the country’s more realistic opportunity may lie in building a disciplined, measurable, and governance-focused applied AI ecosystem.

If India is emphasizing scale and the United States speed and commercial innovation, Azerbaijan’s comparative advantage may be in structured implementation - integrating artificial intelligence into government services, industry, and public administration in ways that deliver practical and measurable efficiency gains.

Note: Etibar Aliyev works on developing and optimizing large language models at Google and Meta, contributes technical expertise to the California AI Audit Bill and the International Association of Algorithmic Auditors (IAAA), and is a senior member of the IEEE. He is also the author of several peer-reviewed studies on AI applications in immersive environments, adaptive narratives, and social network algorithms.