Chatbots have been around for years, providing information and advice - sometimes effectively, sometimes less so - on products, services, and policies. Recent advances in generative AI (GenAI) and large language models (LLMs) are transforming what chatbots can do. These technologies can now deliver richer, more context-aware responses, offering the potential to make public information and services more intuitive, accessible, and inclusive. This can help people find what they need, when they need it.
The Government of Mexico City (Ciudad de México – CDMX) is harnessing these advancements to build a next-generation AI-powered chatbot that will improve how residents and visitors access public and cultural information - especially in the lead-up to the 2026 FIFA World Cup, jointly hosted by Mexico, Canada, and the USA.
Through its Agencia Digital de Innovación Pública (ADIP), CDMX is developing a city-wide digital assistant capable of providing real-time information on topics such as transport, weather, cultural events, restaurants, and museums. By adopting conversational AI and LLM technologies, Mexico City aims to create a trusted, accessible digital interface that shows the full breadth of what the capital has to offer.
But as the ODI often says, without data, there is no AI. The success of any AI system depends on the quality, usability, and ethical foundations of the data that powers it. For chatbots and other GenAI tools, ensuring that data is AI-ready (high-quality, interoperable, well-documented, and responsibly governed) is critical to building public trust and ensuring meaningful outputs.
The ODI’s framework for AI-ready data provides clear guidance and criteria for organisations to assess and improve the datasets they use for AI applications. This includes practical recommendations for data, metadata, and infrastructure design that mitigate bias, improve discoverability, and strengthen interoperability for data publishers. These principles will underpin the CDMX chatbot, ensuring it is built on a foundation of trusted, ethical, and high-quality data.
To do this, the ODI, Faculty AI, and ADIP are working together to deliver a safe, secure, and trustworthy chatbot that demonstrates how AI can be responsibly embedded into public services, representing a next step in the development of open cities.
The project, funded by the UK Foreign, Commonwealth and Development Office (FCDO), combines the ODI’s expertise in data infrastructure, open data, and governance with Faculty’s leadership in AI engineering and responsible AI frameworks. Together, the teams are developing a proof of concept to generate lessons that support:
- Developing a replicable model of responsible, citizen-benefiting AI technology that strengthens Mexico City’s digital innovation capacity.
- AI capacity building within CDMX through training, standards development, and leadership engagement.
- Strengthening the quality and usability of Mexico City’s open data ecosystem to support future AI-enabled public services.
- Producing openly accessible outputs and frameworks that other cities can adopt or adapt.
- Facilitating a city-to-city exchange between Mexico City and London’s digital leadership teams, including at the Greater London Authority, to share best practices and build long-term international cooperation.
This collaboration also signals the UK’s global strengths in responsible AI and data governance, contributing to the UK’s ambition to lead in setting international standards for emerging technologies. By establishing frameworks for trustworthy AI in public services, this project not only benefits citizens and visitors to Mexico City but also lays the groundwork for future cooperation and innovation across Latin America.
Get in touch
To learn more about how the ODI is supporting governments and cities to build trustworthy, AI-ready data infrastructure and digital public services, please contact [email protected]
