BAKU, Azerbaijan, October 20. The Financial Monitoring Service of Azerbaijan is testing artificial intelligence (AI) technologies in pilot mode, Azer Abbasov, Deputy Chairman of the Board of the Financial Monitoring Service, said at the Sustainable Business, Corporate Governance, Risk and Compliance (GRC) conference in Baku on Tuesday, Trend's correspondent reports.
He said that while AI creates new opportunities, its implementation also raises a number of issues related to data security, data quality, and proper training of the system.
According to Abbasov, systems established under the relevant presidential decree and existing databases provide an important foundation for introducing AI in financial monitoring. The technology has been undergoing testing for approximately one to two years.
"Artificial intelligence is a relatively new concept and is evolving rapidly. It analyzes available data and draws logical conclusions. The key issue is ensuring that the data is analyzed correctly and that the resulting conclusions are reliable," he noted.
The deputy chairman emphasized that data protection is particularly important in the Financial Monitoring Service's operations. Its databases contain not only personal information but also tax, banking and commercial secrets, as well as other sensitive data.
According to him, international compliance requirements impose strict restrictions on the storage and processing of such information. Particular attention is therefore paid to server infrastructure and data security.
Abbasov said that entities subject to financial monitoring requirements submit relevant information to the service about transactions they consider suspicious. However, analyzing only this information is not enough to make full and effective use of AI's capabilities.
"If the data consists solely of information submitted by obligated entities and no additional databases are available, the opportunities to enrich that data using artificial intelligence will be limited," he said.
The deputy chairman noted that under the standards of the Financial Action Task Force (FATF), the objective extends beyond identifying suspicious transactions. It is also important to ensure that relevant reports are submitted correctly, improve routine reporting processes and prevent unjustified reporting when there are no reasonable grounds for suspicion.
He emphasized that the effectiveness of AI depends on two main factors: the level of its analytical capabilities and the quality of the data used.
According to Abbasov, the service currently operates an AI system on its local server. The same data is analyzed using both traditional analytical methods and AI, and the results are compared.
He said that further development of the system requires the establishment of relevant behavioral models and continuous AI training. Since the system is not connected to the internet, its improvement must be carried out primarily by the service's own staff.
"To use artificial intelligence effectively, it must be trained, behavioral models must be developed, and its results must be continuously evaluated. At the same time, a cautious approach must be maintained, taking into account the risks that may arise from the introduction of new technologies," Abbasov added.
According to him, when implementing AI in financial monitoring, protecting data confidentiality, ensuring the reliability of results and managing risks must remain key priorities alongside technological innovation.