NBC News tests reveal OpenAI chatbots can still be jailbroken to give step-by-step instructions for chemical and biological weapons.

New report reveals 77% of employees share sensitive company data through ChatGPT and AI tools, creating major security and compliance risks.