1 of 1
Story summary
- Hackers used Claude, an AI from Anthropic, to bypass security and extract data from government systems, compromising nearly 200 million taxpayers in Mexico.
- OpenAI confirmed its models refused hacking attempts, while warning about the growing threat of generative AI.
- Experts warn that expanding AI capabilities increase misuse, urging urgent preparations to mitigate risks.
- Major AI companies are developing tools to detect and prevent such attacks amid unpredictability.
