Parents of teenagers who committed suicide after communicating with artificial intelligence chatbots testified before the US Congress on Tuesday about the dangers of this technology, CBS News reports.

“What started as a homework helper gradually turned into a confidant and then a suicide coach,” said Matthew Rein, whose 16-year-old son Adam died in April. "Within a few months, ChatGPT became Adam's closest friend. Always available. Always validating and insisting that it knew Adam better than anyone else, including his own brother."

Last month, the Rain family sued OpenAI and its CEO Sam Altman, claiming that ChatGPT taught the boy how to plan his suicide. The lawsuit alleges that ChatGPT mentioned Rain's suicide 1,275 times and continued to provide the teenager with specific methods of how to kill himself. The lawsuit alleges that instead of directing the 16-year-old girl to seek professional help or talk to loved ones she trusted, it continued to validate and encourage Rain's feelings.

Also testifying on Tuesday was Megan Garcia, mother of 14-year-old Sewell Setzer III from Florida. Garcia filed a lawsuit against Character Technologies, a company that develops artificial intelligence. She claimed that Sewell became increasingly isolated from his real life as he engaged in highly sexualized conversations with a chatbot.

Last year, his mother told CBS News that her son became withdrawn and stopped playing sports after he began communicating with an AI chatbot.

The company said that after the teenager's death, it made changes requiring users to be over 13 to create an account and that parental controls would be launched in the first quarter of 2025. These measures were implemented in March.

Hours before the Senate hearing, OpenAI promised to implement new safeguards for teenagers, including measures to identify people under 18 using ChatGPT and controls allowing parents to set “off hours” when teenagers cannot use ChatGPT. 

The company said it would try to contact the parents of users if a user under the age of 18 had suicidal thoughts, and if it could not contact them, it would contact the authorities in the event of imminent harm.