Researchers from the independent security platform Hacktron AI reported they used Anthropic's Claude tool to hack OpenAI's ChatGPT model. CBS News MoneyWatch reporter Megan Cerullo reports.