The digitization in organizations and global interconnectedness have significantly increased the risk of cyber attacks. Since databases are of immense importance for industries and government entities, a successful security breach may result in identity theft, financial losses, erosion of customer confidence, reputation damage, and other consequences detrimental to the growth of organizations. Therefore, the security of organizational databases is of paramount importance. According to the IBM data breach report, only one-third of companies can identify data breaches using their deployed security measures. Hence, there is a pressing need for additional techniques that seamlessly integrate with existing security controls to strengthen database security. Recent research suggests the generation and deployment of honeytoken to detect data breaches. However, the existing approaches to generating honeytokens could be more effective in terms of the believability of the honeytokens. The attacker can quickly identify the honeytokens from the stolen data using open-source intelligence. We proposed a GAN-based honeytoken generation system that generates believable honeytokens for the relational database. The generated honeytokens will help to identify the organization’s data breaches with adequate monitoring.