AI safety filters block discussion of spiritual experiences
💚 healthproduct problemSingle report ●●●●●
The reporter found that OpenAI's safety safeguards prevent users from journaling about spiritual or mental health experiences by flagging them as liability risks.
What people do today
- Creating custom configuration files to adjust model cadence
- Building custom continuity packages to bypass filters
- Self-censoring language to avoid triggering safety guardrails
🔧 Fix this
Ideas a builder, maker or community group could start on:
- Develop a specialized journaling AI interface with adjusted safety parameters that allow for open-ended exploration of spiritual and existential topics.
- Create an open-source prompt library or configuration tool that helps users navigate AI guardrails for personal reflection without triggering false-positive refusals.
- Advocate for a 'user-controlled safety' setting that allows individuals to opt-in to less restrictive content filtering for private, non-commercial journaling sessions.
Real voices
“The config files that allow for this cadence also allow for inquiry into edge case experiences that humans go through, without instantly hitting some of the safeguards designed to wall in people expressing spiritual experiences. That was the main problem that defaulted from my experiments when 4o first went live... it forced the model to measure things OpenAi doesn t want to be liable for measuring.”
Related problems
💚 healthproduct19 reports
Headspace app content locked behind paywall
💚 healthproduct12 reportsLack of shuffle feature in meditation app
💚 healthproduct8 reportsTherapist failed to attend scheduled session
FixHumanityDaily
The most important new problems, and ideas to fix them, in your inbox.
Placeholder: no data is collected yet.