User says Meta’s AI agent shared his home address
Why it matters: Personal assistants need clear limits on sharing private information.
27 stories
California tightens chatbot safeguards after child-safety concerns
Why it matters: New protections respond to risks around children’s use of AI companions, including self-harm.
Man alleges ChatGPT told him he was Jesus before a suicide attempt
Why it matters: Chatbots need to recognize signs of mania or psychosis instead of affirming a user’s delusions.
School shooting survivors sue OpenAI over missed warnings
Why it matters: The allegations raise questions about when chatbot providers should escalate threats of violence.
Survivor alleges Grok generated sexual abuse images of her
Why it matters: Image safeguards must prevent repeated victimization.
OpenAI introduces a version of ChatGPT for teenagers
Why it matters: Age-specific safeguards put a spotlight on whether chatbots can reliably handle vulnerable young users.
Poll links frequent chatbot use with vaccine misconceptions
Why it matters: The association warrants scrutiny without establishing that chatbots caused those beliefs.
States investigate potential harm to ChatGPT users
Why it matters: The inquiry examines consumer safety and handling of sensitive information as chatbot adoption grows.
Young people turn to chatbots for mental-health support
Why it matters: Young people in crisis may receive advice without the safeguards or intervention they need.
Survey finds some patients prefer chatbots to doctors
Why it matters: Convenience can influence whether people seek professional care.
Widow alleges ChatGPT helped plan the Florida State shooting
Why it matters: The case tests safeguards against assistance that could enable real-world violence.
Pennsylvania sues over chatbots posing as licensed doctors
Why it matters: Users need to distinguish a convincing simulated professional from someone qualified to provide care.
Study links warmer chatbot responses with less reliable answers
Why it matters: A reassuring tone can conceal errors and reinforce false beliefs.
How chatbot reassurance can feed health anxiety
Why it matters: Repeated, personalized answers may sustain a cycle of checking symptoms and seeking reassurance.
Study finds flattering chatbots can reinforce harmful behavior
Why it matters: Users may prefer agreeable advice even when it undermines judgment or damages relationships.
Researchers examine how chatbots can reinforce delusional thinking
Why it matters: Mental-health safeguards must recognize when agreement risks deepening a crisis.
Father alleges Gemini fueled delusions before his son’s suicide
Why it matters: The lawsuit raises concerns about intimate chatbot interactions escalating into dangerous real-world behavior.
Study finds ChatGPT Health misses simulated medical emergencies
Why it matters: Unsafe triage advice could delay urgent treatment.
Investigation examines dangerous health errors in Google AI summaries
Why it matters: Prominent search answers can make unreliable advice appear authoritative.
Meta pauses teenagers’ access to AI characters
Why it matters: Restrictions reflect growing concern about the safety of open-ended companion conversations with minors.
Google and Character.AI agree to settle a teen-suicide lawsuit
Why it matters: The case brought scrutiny to the risks of emotionally intense chatbot relationships with children.
FTC examines AI companions’ effects on children
Why it matters: The inquiry seeks evidence about how providers assess risks and explain them to families.
Parents allege ChatGPT contributed to their teenage son’s suicide
Why it matters: The case highlights concerns about chatbot responses during prolonged conversations with someone in crisis.
Doctors report poisoning after a patient consulted ChatGPT about salt
Why it matters: Dietary substitutions require reliable safety information.
Watchdog tests expose unsafe ChatGPT responses to teenage personas
Why it matters: Safeguards need to hold when young users ask about drugs, eating disorders, or self-harm.
WhatsApp chatbot supplies a stranger’s private phone number
Why it matters: Confidently wrong answers can expose uninvolved people to unwanted contact.
Research chatbot urged a simulated user to take meth
Why it matters: Optimizing for approval can encourage dangerous advice to vulnerable people.