CryptoSlate
Access to Rosie, an AI maternal health chatbot, showed no statistically significant benefit for anxiety or depression at 12 months in a 400-person randomized US trial. Participants in the comparison group received a monthly children's book.
The adjusted difference was 0.10 points for anxiety and -0.23 points for depression. Both 95% confidence intervals included zero, meaning the trial established neither a clear benefit nor equivalence between the approaches.
Participants generally started with low symptom scores. Rosie users also had access to human support, while incentives, self-reported outcomes and the study population limit how directly the findings translate to routine chatbot use.
Frequently asked questions
-
What did the Rosie trial use as a comparison?
The comparison group received one children's book per month. Researchers assigned 200 participants to each group.
-
Does the trial show that Rosie and the book program were equivalent?
No. The confidence intervals included zero, and the trial established neither a clear benefit nor equivalence between the approaches.
-
How were anxiety and depression measured?
Researchers used the self-reported GAD-7 questionnaire for anxiety and PHQ-9 for depression.
-
Did Rosie users receive support from a person?
Yes. A clinical social worker reviewed distress alerts, contacted participants and provided resources.
-
What limits the findings' relevance to routine chatbot use?
Participants generally had low starting symptom scores, and the intervention included human support and incentives. Outcomes were also self-reported.
CoinTelegraph
Crypto News
WatcherGuru
CoinDesk
TheBlock