From the team
News & Research
Product updates, mental health research, and science-backed insights from the Companion team.
Why We Moved From OpenAI to Anthropic as Our Default AI Vendor
When OpenAI quietly updated its usage policies to permit military and national security applications, we decided to switch. Here is why that decision mattered to us – and to our users.
Read more →Companion Now Speaks Your Language – 100+ Languages Live
Companion now supports over 100 languages natively, powered by Anthropic's Claude for text and OpenAI for voice-to-text – making mental fitness accessible worldwide.
Read more →Introducing the 90-Day Wellbeing Challenge
Our new structured 90-day program guides you through daily check-ins, reflections, and evidence-based exercises to build lasting mental fitness habits.
Read more →WHO: 1 in 4 People Will Face a Mental Health Challenge This Year
The World Health Organization's 2024 Mental Health Atlas underscores a widening global treatment gap – and the urgent need for scalable, accessible support tools.
Read more →Beyond GenAI: Introducing Questioning AI (QuestAI) to Reverse Cognitive Decline
Frequent AI use risks outsourcing not just tasks but the thinking behind them. We propose a pivot from Generative AI to Questioning AI – a framework using critical inquiry and Socratic dialogue to keep humans cognitively in the driver seat.
Read more →The Science of Journaling: How Writing Rewires Your Emotional Brain
Research in affective neuroscience explains why daily journaling – even just five minutes – produces measurable improvements in emotional regulation and stress resilience.
Read more →Study: Generative AI Chatbots Show Significant Potential as Mental Health Tools
A systematic review and meta-analysis of 14 randomised controlled trials finds that generative AI chatbots can meaningfully reduce mental health issues such as depression and anxiety.
Read more →The Evaluation Framework Behind Companion: HAICEF Explained
A 2025 JMIR systematic review standardised how healthcare AI chatbots should be evaluated. Here is how Companion applies the HAICEF framework across Safety, Fairness, Trustworthiness, and Usefulness.
Read more →Why People Open Up More to an AI Than to a Human
Counterintuitively, studies show users disclose more personal and emotionally sensitive information to AI agents than to human therapists. Here's the psychology behind it.
Read more →