SAFE KIDS Act

Key Points

• AI chatbot companies must verify whether users are under 18. If they cannot confirm a user is an adult, the company must treat them as a child and apply all safety protections by default. Companies can use age estimation technology or rely on signals from operating systems and app stores.
• The bill bans all advertising to children using AI chatbots, including product placement in conversations. It also makes it illegal to sell or share a child's personal information without verified parental consent, and companies cannot force kids to hand over extra data just to use the chatbot.
• AI chatbots cannot pretend to be human, encourage emotional dependence, promote self-harm, show graphic violence or sexual content, or discourage children from talking to parents and trusted adults. Companies that knowingly or recklessly allow these behaviors face penalties.
• Parents get a suite of controls including time limits, feature restrictions, memory controls, and the ability to block AI training on their child's data. Parents also receive immediate alerts if the AI detects their child may be at risk of suicide or self-harm.
• Every AI chatbot provider must pay for an annual independent safety audit. The FTC can impose civil penalties of $1,000 per violation per user for safety failures, and up to $10,000 per user for willful violations involving child sexual exploitation or submitting false information.
• The law would take effect 180 days after enactment, and the FTC must set up a public incident reporting system and an online resource linking to each chatbot's child safety policy within that same timeframe.