Meta adjusts AI behavior amid lawsuits over data and invasive
Meta is revising its AI prompts and facing lawsuits over alleged unauthorized use of photos for AI training and face-recognition systems.
Meta revises AI prompts after invasive questioning backlash
Meta is changing how its AI suggests prompts after a viral video showed it asking a user to identify her child in a video. The AI had also pieced together personal details about her family from past posts. Meta spokesperson Dina El-Kassaby said the company "missed the mark" and that the feature "never should have prompted the individual with questions like that." The AI only surfaces content the user already has access to, according to Meta.
Lawsuit alleges unauthorized use of photos for AI training
A proposed class action lawsuit filed in federal court alleges Meta illegally used Facebook and Instagram photos to train its AI image-generation models (Emu, Muse Image) and its unreleased "NameTag" face-recognition system. The plaintiffs claim Meta violated Illinois and California privacy laws by extracting biometric data without consent. Meta denies the allegations, stating it has been transparent about AI training practices.
NameTag face-recognition system under scrutiny
Code for NameTag, an unreleased feature for Meta's smart glasses, was found embedded in an AI companion app downloaded over 50 million times. The system was designed to match faces captured by glasses against a local faceprint database, which could receive updates from Meta. The lawsuit suggests the faceprints may derive from user photos on Meta's platforms. Meta removed the code after media reports but defended its development as exploratory.
Previous biometric data settlements haunt Meta
Meta has faced prior legal action over biometric data handling. In 2020, it paid $650 million to settle an Illinois class action over its earlier face-recognition system. In 2024, it agreed to pay $1.4 billion to resolve separate allegations of unlawfully collecting biometric data. The current lawsuit seeks $5,000 per intentional or reckless violation under Illinois law, or $1,000 per negligent violation, as well as additional damages under California law, with a potential class size in the millions.
AI training practices draw legal and ethical concerns
The lawsuit highlights broader concerns about how Meta trains its AI systems. Meta has acknowledged using public Facebook and Instagram images to train models like Emu, calling it a "data advantage." Muse Image faced criticism for allowing image generation based on public Instagram accounts, a feature Meta quickly removed. Plaintiffs argue users shouldn't have to worry about biometric misuse simply for posting photos.
What creators should consider
Media creators using Meta platforms should review their privacy settings and understand how their content might train AI systems. Those concerned about biometric data use can explore opt-out options where available. The legal outcomes could set precedents for how social media platforms use user-generated content in AI development.
