WASHINGTON — Sen. Bernie Sanders, an independent from Vermont, and Rep. Greg Casar, a Texas Democrat, introduced legislation Wednesday that would ban the development of artificial “superintelligence” and pause advanced AI work until the federal government can establish stronger oversight. The Ban Artificial Superintelligence Act would create a Department of Artificial Intelligence led by a Senate-confirmed secretary. The proposed agency would monitor AI capabilities, identify potential risks and enforce limits on the most advanced systems. Casar said the bill would prohibit “recursive self-improvement,” in which an AI system develops more capable versions of itself. It also would bar models from designing nuclear or biological weapons and prohibit systems that attempt to evade human supervision or deceive people. The legislation would impose severe consequences on companies and individuals that try to bypass federal oversight. Companies could face what the lawmakers described as a “corporate death penalty,” while individuals could receive prison terms of up to 20 years — a punishment comparable to the penalty for illegally developing nuclear weapons. Sanders and Casar have been pressing for a pause on advanced AI development amid warnings from technology insiders and concerns about cyberattacks. Sanders said lawmakers should not wait for a catastrophic event before acting. “In my view, we cannot simply sit by idly and wait for advanced AI to cause the next 9/11 or Chernobyl or a global pandemic or a financial crisis before we take action to protect the American people and, of course, the people of the world,” Sanders told reporters. The proposal arrives as Congress considers narrower AI measures, including a bipartisan whistleblower bill backed by Senate Judiciary Chair Chuck Grassley, an Iowa Republican. That bill would protect workers who disclose information about risks in AI development but would not establish the sweeping regulatory system Sanders and Casar are seeking. The two lawmakers expressed confidence that both chambers could take up their bill in the coming weeks. But the proposal faces a difficult path: Democratic leaders have focused on more limited regulatory steps, while some Republican leaders have rejected the need for new AI safety legislation altogether. Casar said public pressure for safeguards could make the bill a practical starting point for Congress. Whether lawmakers can agree on rules before AI systems become more capable remains an open question, particularly as private companies continue to drive development of the technology.