What should they have done instead?

I agree. This also highlights the peril of taking a moral purity approach to judging donations. Often, when ppl take money from a morally fraught person or organization, it just amounts to valuable resource redistribution.

Find your local No Kings peaceful protest: www.nokings.org?SQF_SOURCE=h...

Congrats, Hahrie! This is much deserved and wonderful to see.

Also, you can't be simultaneously upset that the ML classifier gets things wrong and be upset at the thought of training on user data, because the two are mutually exclusive. The only way to get more accurate ML models is to retrain with real-world examples.

Paper: osf.io/preprints/ps...
OSF/prereg: osf.io/g59fm/
All chatbots linked above!

Major kudos to brilliant post-doc Jonas Schoene, who led this project!

Thanks to co-authors: Johannes Eichstaedt, Aadesh Salecha, Sonja Lyubomirsky

And to: @stanfordhai.bsky.social
@stanfordpacs.bsky.social
@pascl-stanford.bsky.social
@uclaanderson.bsky.social

We find high public interest in using validated well-being chatbots.

In a nationally rep survey (N=3,056), about 1/2 of U.S. adults said they’d use empirically validated well-being chatbots. About 1/4 have already used #generativeAI chatbots for well-being.

Importantly, results generalized to key subpopulations, including those with high baseline levels of anxiety or depressed mood.

Linguistic analyses showed users had diff conversations with each chatbot:

Savoring → Sensory perceptions & positive emotions
Gratitude → Social interactions & positive emotions
Meaning → Reflection on others & cognitive processes
Hero’s journey → Achievement & cognitive processes