The Infocomm Media Development Authority (IMDA) has on Monday published new voluntary guidelines to help consumers navigate the use of Generative AI (GenAI) Chatbots.
As GenAI chatbots become embedded in everyday life, transparency is crucial as it allows users to understand what these applications can and cannot do, and how key risks are managed, IMDA said in a statement.
The Transparency Guidelines for GenAI Chatbots, amongst the first of its kind in the world, aims to standardize how information is shared with the public, while creating a trusted ecosystem where users can engage GenAI with confidence, built on a foundation of transparency.
The guidelines set out how GenAI chatbot deployers can provide meaningful transparency to consumers.
They cover what information consumers may find relevant and how deployers can share it clearly, accessibly, and in a timely manner.
The guidelines aim to enable informed use, standardize disclosure and enhance accountability.
It is noted that when users understand an application’s capabilities and limitations, they can make better decisions, verify outputs where appropriate, and recognize when use may be inappropriate.
Meanwhile, a consistent framework helps users compare applications and find relevant information quickly, rather than navigating multiple documentation.
By stating their own safety commitments clearly, deployers have an opportunity to set a public standard for their chatbot’s safety and reliability, against which users can better understand and assess the service.
At the heart of the guidelines is a chatbot info card. This can take different forms, such as a dedicated information page or a disclosure document.
Like a nutrition label or medicine label, it gives users a single, consolidated place to find essential information about the chatbot they are using.
The info card is built around three straightforward principles – Relevance, Accessibility and Timeliness.
Firstly, the info card should answer the questions users actually care about: what the chatbot can and cannot do, how reliable and safe it is, how their data is handled, and how to raise concerns or report issues.
Secondly, the information should be written in plain language, easy to navigate, and easy to find.
Thirdly, the info card should remain up to date. Deployers are encouraged to publish it when the chatbot is made available to users and to update it when significant changes affect the chatbot’s capabilities, risks or safety policies.
The guidelines were developed with input from industry, Government agencies and consumers.
IMDA said it engaged a wide range of stakeholders, including major technology companies, enterprises with consumer-facing chatbots, public agencies and local chatbot users, to better understand current transparency practices, practical implementation considerations, and the information consumers would find useful.
The guidelines seek to balance these perspectives by providing a practical framework to improve transparency in a way that is useful to consumers and realistic for chatbot deployers to implement.
It is noted that organizations across the AI ecosystem have expressed support for greater transparency in AI deployments.
Several organizations have also indicated that, over the next six to 12 months, they intend to take reference from the guidelines in strengthening the transparency practices for their public-facing chatbots.
These organizations recognize that clear disclosure can help foster trust, improve user understanding and encourage responsible adoption of AI technologies.
These include global AI companies such as Google (for the Gemini app), Meta as well as major AI deployers in Singapore, such as DBS, OCBC, Singapore Airlines and Synapxe.
Sector regulators also recognize this effort. As AI adoption continues to accelerate in Singapore’s financial and healthcare sectors, the Monetary Authority of Singapore and the Ministry of Health have issued sector-specific guidance on responsible use of AI, namely the Proposed Guidelines on Artificial Intelligence Risk Management and the updated Artificial Intelligence in Healthcare Guidelines respectively, both of which emphasize transparency as an important principle.
IMDA said its transparency guidelines complements these sector-specific guidance, particularly where organizations deploy GenAI chatbots.
The Singapore public sector will also lead by example. Public sector agencies, including the National Library Board and Health Promotion Board, plan to take reference from the guidelines when providing information for their own public-facing chatbots.
Taken together, the broad support from global AI companies, industry deployers, sector regulators and the public sector reflect a shared commitment to promoting responsible AI adoption and strengthening public confidence in AI innovation through greater transparency, said the statement.
The guidelines center on GenAI chatbots as a starting point, given their scale, high consumer touchpoint, and growing societal concerns around data privacy, the safety of minors, and risks to mentally vulnerable users.
Nonetheless, deployers can use the guidelines as a reference point when providing transparency for other GenAI applications. Sectoral regulators can also build on them to develop more specific transparency guidelines for applications in their sectors.
IMDA highlighted that these transparency guidelines mark a significant step in Singapore’s efforts to build a trusted AI ecosystem.
They provide practical guidance for chatbot deployers to communicate their capabilities, limitations and safety policies to users, while preserving flexibility for them to implement transparency measures in ways that fit their services and their users.
As AI expands into sectors such as healthcare, education, and financial services, IMDA said it will work with sectoral regulators and other partners to extend and adapt these principles to address the unique transparency demands of different domains.
It noted Singapore remains committed to contributing to global efforts on AI transparency, working alongside international organizations and standards bodies to shift the conversation from model-level to application-level transparency — with the shared goal of an AI ecosystem that is innovative and worthy of public trust.
This is in line with Singapore’s practical and balanced approach to AI Governance, where guardrails support trust while providing space for innovation, it added.
IMDA launches Singapore’s first Digital Twin Playbook for Enterprises