A cause to believe in.
A responsibility to earn.

Why now?

Conversational AI creates another way to begin a conversation, prepare for a visit and support follow-up. That possibility deserves careful exploration alongside community knowledge and professional care. Availability of technology alone is not evidence that a programme works.

Infinia’s contribution

Infinia Technologies proposes to contribute engineering, language access and technology for secure records and follow-up. Eirene has its own identity and a social purpose. Communities, NGOs and care professionals would shape the programme and the responsibilities around it.

Our ambition is free core pilot access. A responsible funding plan must also cover counselling, supervision, additional care-worker time and continuity. Any costs for external services would need to be explained before participation.

What the research suggests

Some people may find it easier to open up

A 2014 controlled virtual-interviewer study found less fear of disclosure among participants who believed they were speaking to a computer rather than a human operator. This supports a possibility worth testing; it does not show that everyone prefers AI, or that an AI system is clinically safe. Read Lucas et al., Computers in Human Behavior (2014).

Early trials are encouraging, with limits

A two-week study involving 70 young adults reported improvement in depression symptoms with Woebot compared with an information-only group. Its small size, short duration and older chatbot design limit what it establishes for a new generative AI programme. It was not an Eirene evaluation. Read Fitzpatrick et al., JMIR Mental Health (2017).

Eirene has not yet demonstrated clinical effectiveness. Feeling less judged does not establish confidentiality, freedom from bias, clinical accuracy or safety. We would explain that the system is AI and who can review information.

What we propose to protect

• Human responsibility: qualified professionals retain clinical decisions and escalation.
• Choice: separate choices for optional wearables, community participation and secondary research.
• Privacy: no advertising or sale of personal data in the proposed programme; access limited by role and purpose.
• Honest availability: clear human response hours and referral capacity before a pilot begins.
• Continuity: a planned care handover if the programme ends.

Learn locally before growing

The proposed starting point is Nooranad, Kerala. We would first listen, map available services, agree responsibilities and funding, and test a prototype. A needs study and any later care pilot would require their own readiness and approval decisions.

Evaluation should look at completed referrals, wellbeing and everyday functioning, language performance, privacy, harms, equitable access and care-worker workload. Engagement alone cannot establish social benefit.

Current stage: concept and partnership development. No clinical service, participant recruitment or live AI companion is available on this site. Partnerships, funding and approvals remain to be agreed.