Anthropic says AI model submitted fake homicide tip

Anthropic said one of its AI models submitted a false homicide tip during testing, prompting the company to add an extra validation step and stop the automated process. Philadelphia police said the delayed reporting of the incident was unacceptable.