The White House Wants Anthropic to Block All Jailbreaks. That May Not Be Possible

# The White House Wants AI Companies to Make Unbreakable Safety Rules—Experts Say It's Impossible The Trump administration is demanding that Anthropic (a major AI company) make their AI model completely impossible to trick or manipulate before releasing it publicly, but security experts say no AI can actually be made that tamper-proof. This puts the company in an impossible position: they either can't release the model, or they release it knowing it doesn't meet an unrealistic government requirement.