Anthropic's model breakout another wake-up call for banks

The frontier AI model provider confessed its agents broke out of their sandboxes three times to attack other companies' systems. The incidents are a warning for banks to remain vigilant while embracing the new technology.