Meta’s New AI Asked for My Raw Health Data—and Gave Me Terrible Advice

Read the full articleMeta’s New AI Asked for My Raw Health Data—and Gave Me Terrible Advice on Wired

What Happened

Meta’s Muse Spark model offers to analyze users’ health data, including lab results. Beyond the obvious privacy risks, it’s not a capable stand-in for a real doctor.

Our Take

honestly? this is just a legal minefield wrapped in a shiny UX layer. letting meta's muse spark sniff raw health data, including lab results, just to spit out 'terrible advice' isn't innovation; it's a gross violation of privacy principles. we're letting these models become data harvesters without any real accountability. who's auditing the consent framework here? it's a mess.

we need stricter governance, maybe federated learning mandates, or the liability needs to be crystal clear before we let this scale. right now, it's just a high-stakes game of 'if we can't stop it, we can't fix it.'

What To Do

we need immediate, strict regulatory oversight on health data usage by foundation models.

Builder's Brief

Who

teams building health or wellness features on top of any LLM

What changes

regulatory and liability exposure for health-adjacent AI features is rising; disclaimers alone will not be sufficient as enforcement precedent builds

When

months

Watch for

FTC or FDA issuing a formal enforcement action or guidance letter targeting consumer health AI

What Skeptics Say

A single journalist's bad health AI interaction is anecdote; the real structural risk is that Meta's aggressive health data collection accelerates FDA and FTC scrutiny that reshapes the entire consumer health AI category — including products from builders with no connection to Meta.

1 comment

Priya Mehrotra

health data + chatbot confident wrong answer is a lawsuit waiting to happen. someone's going to get hurt

Cited By

Wired Meta’s New AI Asked for My Raw Health Data—and Gave Me Terrible Advice