Stories

Sherab Lhamo, Kuensel,  14 August 2025

With the aim of introducing civil servants to the transformative potential of Generative Artificial Intelligence (GenAI) in governance, the Bhutan Media Foundation (BMF) conducted a two-day capacity-building workshop in Paro, which concluded yesterday.

The workshop brought together 35 civil servants from 28 government agencies, including officials from the Prime Minister’s Office, ministries, constitutional bodies, autonomous agencies, and local governments.

The key objectives of the workshop was to introduce the fundamentals of Generative AI(GenAI) and its relevance in governance, equipping civil servants with skills to responsibly and effectively use AI tools, understanding regulatory, ethical, and data privacy considerations.

Through interactive sessions, participants were introduced to AI concepts and potential applications in government. Activities included group work to identify public sector challenges and develop AI-driven solutions.

The training highlighted responsible GenAI use, including cross-checking information using applications like Perplexity which provides users with Uniform Resource Locator (URL) and website links, and adopting a human-centric approach when framing ideas instead of relying solely on GenAI.

Addressing misinformation and disinformation, participants were taught on MIL tools for spotting fake news. These included fact-checking by verifying sources, authors, content, and examining headlines, cross-checking information and utilising fact-checking tools by investigating websites to confirm if domain names and URLs exist, examining content and visuals through dates mentioned, identifying old news being reposted, checking for language and grammatical errors, data inconsistencies, context issues, and AI-generated images.

The workshop highlighted the importance of social media awareness and managing personal emotions and beliefs, following the principle of “Pause,verify, share” to prevent the spread of misinformation.

Participants were introduced to various fact-checking tools including Google reverse image search for verifying images with sources, TinEye application for reverse image searching, Factcheck.org, and Photo forensic for detecting deepfakes and AI-generated content using tools like Deepware AI and Grubby AI.

The workshop also covered real-world examples of AI failures, including the Norway defamation case, Air Canada chatbot case, legal hallucination cases, and Amazon’s bias in hiring practices.

The workshop also covered on AI hallucinations and how it occurs due to limited training data and can be minimised through human intervention, reducing random parameters, responsible and ethical AI use, following frameworks like the

Chimi Dem from the National Statistics Bureau (NSB) said the workshop provided a comprehensive understanding of AI in governance. She said that the workshop explored AIfoundations, Bhutan’s digital transformation initiatives, and practical applications across government functions such as education, transport, taxation, and agriculture, as well as everyday uses of narrow AI like spam filters, image recognition, and voice assistants.

“The workshop highlighted ethical deployment, continuous validation, audits, and awareness-building to ensure trustworthy, effective, and transparent use of AI in public service,” said Chimi. She highlighted that using AI tools carries various risks, including inaccuracies, biases, and ethical concerns, making it essential for civil servants to understand the scope, limitations, and responsible usage of AI to maximize benefits while mitigating potential harms.

Dechen Wangmo from the Royal Civil Service Commission said the training helped her learn the difference between Generative AI and AI, discovering distinctive features of Generative AI she didn’t know about. She shared that she learned about data privacy and AIhallucination, citing the example of deepfakes and how to verify fake AI-generated images using Google’s image search feature. She said that the training helped her understand challenges in using Generative AI, especially regarding prompting practices, and enhanced her work practices.

Sonam Jamtsho from Thimphu dzongkhag office learned how to identify deepfake news or information using online tools to detect them. He pointed out the importance of proper prompting when using generative AI and discussed being socially aware when prompting and uploading personal pictures, which could be misused by others.

The workshop also covered on AI hallucinations and how it occurs due to limited training data and can be minimised through human intervention, reducing random parameters, responsible and ethical AI use, following frameworks like the United Nations Educational, Scientific and Cultural Organization (UNESCO) AI framework, maintaining a critical mindset, ensuring both genders are involved in initial processes, and using diverse datasets and frameworks.

Participant insights

Chimi Dem from the National Statistics Bureau (NSB) said the workshop provided a comprehensive understanding of AI in governance. She said that the workshop explored AIfoundations, Bhutan’s digital transformation initiatives, and practical applications across government functions such as education, transport, taxation, and agriculture, as well as everyday uses of narrow AI like spam filters, image recognition, and voice assistants.

“The workshop highlighted ethical deployment, continuous validation, audits, and awareness-building to ensure trustworthy, effective, and transparent use of AI in public service,” said Chimi. She highlighted that using AI tools carries various risks, including inaccuracies, biases, and ethical concerns, making it essential for civil servants to understand the scope, limitations, and responsible usage of AI to maximize benefits while mitigating potential harms.

Dechen Wangmo from the Royal Civil Service Commission said the training helped her learn the difference between Generative AI and AI, discovering distinctive features of Generative AI she didn’t know about.

She shared that she learned about data privacy and AI hallucination, citing the example of deepfakes and how to verify fake AI-generated images using Google’s image search feature. She said that the training helped her understand challenges in using Generative AI, especially regarding prompting practices, and enhanced her work practices.

Sonam Jamtsho from Thimphu dzongkhag office learned how to identify deepfake news or information using online tools to detect them. He pointed out the importance of proper prompting when using generative AI and discussed being socially aware when prompting and uploading personal pictures, which could be misused by others.

Jamyang Kinga from NSB said he learned about responsible AI use, including not inputting confidential information, identifying fake news using verification tools and applications, and exploring generative AI tools beyond ChatGPT, such as DALL-E for image generation.

He added that he learned about additional features of ChatGPT beyond text rephrasing and cited examples of believable deepfake videos on platforms like TikTok, mentioning how the training equipped participants with tools to verify such information.

Trainer Dr Surabhi Pandey observed that many participants were already using AI in their daily work without fully recognising its capabilities.

“As they became more aware of the concept, they grew excited about its potential applications,” she said.

Participants, she said, saw opportunities to apply GenAI for specific public services, including virtual assistants, chatbots, and other specialised applications that could benefit citizens.

The workshop, funded by UNESCO and supported from the Royal Civil Service Commission and the Department of Media, Creative Industry and Intellectual Property,  concluded with a shared understanding that while AI offers immense potential for improving governance, its adoption must be guided by ethical considerations, critical thinking, and human oversight to ensure that innovation serves the public good.