Privacy and Compliance Review: AI on Patient Data

[iprasannamb]

AI on Patient (Marginalized people) Data

Beehive integrates Google Gemini (google.generativeai API) for AI-based analysis. Because Beehive processes patient data, sending images to a third-party AI service constitutes a third-party disclosure and may raise privacy, security, and regulatory compliance concerns, including potential data retention or model training outside our control.

As a precaution, I have disabled AI analysis by default and made this explicit in PR #437.

Decision Required

We need to determine whether this AI feature should:

• Be removed entirely
• Remain disabled by default with strict safeguards

@pradeeban , please let me know your thoughts on this.

Thank you.

[pradeeban]

Oh my god, thanks for this. Merged.

[iprasannamb]

@pradeeban ,

Could you please share your thoughts on the points I raised regarding Al usage on patient data when you have time?

Thank you.

[pradeeban]

Disabling is sufficient for now. So, #437 is fine.
Honestly, David and I are against all forms of censorship on Beehive. @mdxabu knows that.
Some applicants implemented the feature and I merged it for now since the implementation looks solid. But it is not a feature that we need in Beehive. Beehive will be used by a small, well-known set of individuals - around 10 to 25 people in Anchorage. We want no form of censorship there actually.

[iprasannamb]

Thanks for the clarification .

[pradeeban]

Beehive is sadly bloated with all the features we do not need. But perhaps, Beehive may have an audience we did not anticipate initially. Like a local Instagram!