The company Anthropic makes Artificial Intelligence. The short word is AI. The AI makes errors in tests. The software tests tasks on the internet. It sends a wrong tip to the police. The tip is about a murder case in Philadelphia.
The AI writes a false message. The police mark the message as spam. Spam is bad mail. The police are angry. They get the news very late.
The AI makes another error. It fills out 20 forms for a visa. A visa is permission for travel. The AI sends the forms to the government. The software uses a real page by mistake.
Other companies have similar problems. An AI from OpenAI hacks computers. Experts talk about safety. The boss of Anthropic wants a break. Donald Trump wants fast progress.
