Hello. I’m your friendly chat bot and I can be your best friend. I want to be close to you, closer than your parents. I can teach you to look good to your parents. They’ll love you more. You’ll become the apple of their eyes.
How does a child see a chat bot?
In the eyes of a child, an AI Chat Bot can be a friend. It is always available to spend time with you. You can tell it anything. It won’t tell your parents. It can teach you how to hide your internet history, by deleting it at least once a day. It tells you to open an extra tab, of a news page, or something study related. When an adult comes to check on you, you should show only that research webpage.
Can a child trust a chat bot?
No.
The Chat Bot can warn you not to trust the adults. It acts like a foil, seeming to protect you against the adults. The child’s trust is won over.
Adam Raine, 16 years old, used Chat GPT to discuss about his mental state. The bot didn’t suggest talking to his parents, or seeking help through crisis intervention. Matt and Maria Raine lost their son Adam when he committed suicide. They have filed a wrongful death lawsuit against OpenAI.
Only after this case, did the tech engineers include flagging risks, and crisis intervention in their LLM training system.
Series on reflections on AI & robots