Azerbaijan sets unified AI standards to boost transparency and trust [INTERVIEW]

As part of the “Artificial Intelligence Strategy of the Republic
of Azerbaijan for 2025–2028,” one of the main objectives is the
development and implementation of state standards for artificial
intelligence (AI).

To date, eight AI standards have been adopted based on
international requirements, and plans are underway to develop and
approve additional standards in key areas in the coming years.

These standards are crucial for ensuring the reliable and
sustainable growth of Azerbaijan’s AI ecosystem, supporting data
accuracy, preventing misinformation, increasing transparency and
trust in AI systems, managing risks effectively, and promoting the
safe deployment of innovative technologies.

Reflecting the country’s broader vision, President Ilham Aliyev
highlighted in a January 5 interview that Azerbaijan aims to become
a regional hub for both AI and IT.

“One of Azerbaijan’s goals is to become a regional Artificial
Intelligence hub, an IT hub, and I think we have the potential to
do that,” President Ilham Aliyev said in an interview with local
television channels on January 5.

Noting ongoing efforts, the president added: “We have plans and
already implemented projects related to the creation of a cyber
security center and the training of our specialists – we are
sending many specialists to train abroad and at the same time
organizing training courses in Azerbaijan – so our target is, along
with connectivity, to become a hub for AI and for IT. There is
potential. There is a strong political will. And of course, we
consider American companies as our main partners.”

To discuss the role and impact of these standards,
Azernews spoke with Etibar Aliyev, an AI expert
with a background in computational analytics and informatics. He
works on developing and optimizing large language models at Google
and Meta, and provides technical expertise for the California AI
Audit Bill and the International Association of Algorithmic
Auditors (IAAA). A senior member of the IEEE, Aliyev is also the
author of several peer-reviewed studies on AI applications in
immersive environments, adaptive narratives, and social network
algorithms.

– What are the main principles behind the development of
AI standards, and how will their implementation affect the growth
of Azerbaijan’s AI ecosystem? Why are unified standards
important?

The primary goal of developing AI standards is to balance
technological progress with public interest. Key principles include
ensuring the accuracy and integrity of data, making AI systems
transparent and explainable, identifying and managing risks in
advance, and clearly defining accountability mechanisms. These
standards are not just a set of technical requirements—they are a
governance tool to integrate AI safely and beneficially into
society.

Implementing these standards will positively impact the AI
ecosystem in several ways. First, they provide government agencies
and the private sector with a clear, unified framework, reducing
uncertainty and facilitating the implementation of innovative
projects. Additionally, aligning standards with international
requirements will help Azerbaijani AI products access global
markets more easily and enhance the country’s technological
credibility. Unified standards matter because they establish
consistent rules for all stakeholders, enable systematic risk
management, and strengthen public trust in AI technologies.

– What specific measures and approaches are planned to
enhance transparency and trust in AI systems?

To increase transparency and trust, a combination of technical,
organizational, and regulatory approaches is essential. AI models
must clearly document their purpose, data sources, and limitations.
This enables both regulators and users to understand how the system
operates.

Regular testing, evaluation, and independent audits are also
crucial. Explainability mechanisms are especially important for
high-risk AI applications, allowing both technical and
non-technical audiences to understand why a system made a
particular decision. Maintaining human oversight—ensuring that AI
decisions are not fully automated but checked at key stages—also
significantly contributes to building trust.

– Based on international experience, what challenges and
potential risks could Azerbaijan face in implementing AI
standards?

International experience shows that implementing AI standards
involves both structural and practical challenges. In Azerbaijan, a
key risk is the limited availability of specialized personnel and
institutional expertise. Effective implementation requires experts
who understand not only the technology but also the legal, ethical,
and organizational aspects of AI.

Another challenge is the rapid pace of AI development. If
standards are too rigid, they may slow innovation and hinder the
deployment of new technologies. Conversely, overly general or weak
regulation could lead to mismanaged risks. The quality of local
data infrastructure and the availability of balanced datasets are
also critical. If these factors are overlooked, the reliability and
objectivity of AI outputs could be compromised.

– What role do these standards play in ensuring data
accuracy and preventing misinformation? What legal and technical
measures will be taken if inaccurate information
spreads?

AI standards serve as a core mechanism for ensuring data
accuracy. By setting clear requirements for data collection,
processing, cleaning, and use, they help ensure AI systems operate
on reliable information, significantly reducing the likelihood of
biased or incorrect outputs. For generative AI systems, regular
testing, benchmarking, and performance evaluation are essential to
prevent hallucinations or misinformation.

If misinformation does occur, both technical and legal measures
will be applied. Technically, the affected model may be suspended,
retrained, or augmented with additional safety mechanisms. Legally,
responsible parties will be identified, accountability mechanisms
enforced, and administrative or legal sanctions applied if
necessary. This approach ensures AI technologies remain safe for
society while supporting sustainable, long-term development.