Assam ASHAs help design multilingual AI chatbot for cervical cancer screening

Around 100 Accredited Social Health Activists (ASHAs) from Assam took part in a co-design workshop in Guwahati to help develop an artificial intelligence-enabled training platform and multilingual voice chatbot focused on cervical cancer screening and community counselling.

The workshop, organised by The George Institute for Global Health India in collaboration with the Cachar Cancer Hospital and Research Centre (CCHRC), brought together ASHAs, doctors, trainers, programme coordinators, researchers and technology experts.

The ASHAs shared their experiences of helping women access cervical cancer screening and outlined the difficulties they face during household and community visits. They also discussed the kind of digital tools that could help them answer questions, counsel women and access reliable health information while working in the field.

The workshop forms part of a three-year pilot study, ‘Leveraging Artificial Intelligence for Strengthening Cervical Cancer Screening Through Capacity Building’, funded by the Indian Council of Medical Research (ICMR). The project is being carried out between November 2025 and November 2028.

A key exercise involved 155 commonly asked questions related to cervical cancer and screening. ASHAs used bindis to vote for the questions they encounter most often from women, husbands, family members and others in their communities.

The exercise showed a gap between questions considered clinically important and those that ASHAs frequently encounter during field visits. Researchers said the responses would help build the knowledge base of the proposed chatbot, ensuring it addresses both priority health information and practical concerns raised at the community level.

The workshop also examined how ASHAs use digital technology. Younger and middle-aged participants said they routinely use mobile phones for reporting, communication and accessing information, while some older workers reported difficulties with digital tools and language. However, several said they were willing to improve their digital skills.

Participants identified counselling women, explaining test results, addressing concerns related to pregnancy and childbirth and making appropriate referrals as key areas where additional training could help.

They also suggested features such as short audio-visual lessons, frequently asked questions, reminders, quizzes and progress tracking. ASHAs stressed the need for voice-based interaction in Bangla, simple-language content and search functions that allow them to ask questions in the way they would naturally frame them.

One ASHA worker said the workshop offered a different approach to learning about cervical cancer.

"I am very happy to attend this workshop. This is the first time I have attended something like this. I have attended many meetings before, but this was very different."

She added that the bindi-based exercise made the session more interactive and helped participants understand which information was important when discussing cervical cancer and screening.

Dr Anita Gadgil, principal investigator of the project, senior research fellow at The George Institute for Global Health and co-director of the WHO Collaboration Centre for Emergency Critical and Operative Care, said the technology would be developed around the experiences of ASHAs.

"Technology for ASHAs should be designed with ASHAs. Their experience of explaining screening, responding to concerns and finding answers in real time will guide the language, content and functions of this AI platform," Gadgil said.

She added that the objective was to provide clinically validated support that fits into the existing work of ASHAs rather than create an additional digital burden.

The proposed platform will supplement existing training and provide ASHAs with clinically verified information while accounting for challenges such as limited internet connectivity and the limited time available during household visits.

Researchers will next develop and test the AI-based learning platform and chatbot. The training material and questions will undergo clinical review and repeated testing with ASHAs before the technology is rolled out.

The intervention will subsequently be evaluated through a two-arm cluster-randomised controlled trial involving 20 clusters and around 110 ASHAs. The study will examine the technology's usability, acceptance and usage patterns, as well as its potential for wider implementation. Researchers will also assess changes in ASHAs' knowledge, attitudes and practices.

The chatbot is intended to support ASHAs with training and field-based information, but will not replace clinical assessment, diagnosis or existing referral systems.

With cervical cancer continuing to pose a significant public health challenge in India, researchers say improving awareness and access to screening remains important for early detection. ASHAs play a key role in connecting women in rural communities with information and healthcare services.