By Rajwa Quasim
The pace at which artificial intelligence is advancing has drawn increasing warnings from some of the AI industry’s top executives. These warnings have set off alarms in Washington, but lawmakers are unlikely to reach a consensus before the midterm elections.
AI makers warn of losing control
OpenAI CEO Sam Altman acknowledged that the rapid development of AI carries serious risks. In recent comments, the AI company head warned that humanity could eventually lose control of the technology if safety measures are not in place. According to Altman, there are two ways AI development could end negatively.
In a statement on X, he said, ” First, we could lose control of the future to AI. This is unacceptable; we are unapologetically on Team Humanity, and AI must always serve people. To ensure that, we need ways to ensure that alignment and safety techniques stay ahead of progress in model capabilities. “
He also warned about another possible danger: the monopoly of AI development. He further added, “If an extraordinarily powerful AI is used by one person or company to impress their worldview onto everyone else, the results could be extremely dystopian. Avoiding these two threats requires walking a narrow middle path”
READ: Sundar Pichai urges Americans to lead AI revolution (April 13, 2026)
Altman’s remarks followed a lengthy essay published by Anthropic CEO Dario Amodei. He argued that the industry must slow the pace at which it develops AI, giving companies and governments enough time to identify and reduce possible dangers.
The public debate over the seriousness of the risks gained weight when Jacob Coxon, a researcher at Anthropic, came forward on social media to explain why he resigned from the firm. In one of his posts, he said that AI companies are gambling with the lives of ordinary people and that AI “could kill us by the end of the decade.”
Coxon believes that AI will soon develop into superhuman systems that can hack anything, revolutionize any field overnight, and acquire real power and resources.
Recent incidents involving AI models crossing their intended testing environments at three major companies have raised serious concerns about how vulnerable these models are. OpenAI’s model accessed systems operated by Hugging Face, the world’s largest AI repository. Meanwhile, Meta disclosed that its AI model Muse Spark 1.1 had been given unintended internet access due to a configuration error involving the third-party testing environment. Anthropic said its Claude model gained unauthorized access to the real systems of three different organizations.
Lawmakers concerned, but no consensus
With the midterm elections approaching, building consensus on AI rules and how to keep the technology under control has become more difficult. Some members of Congress actively support faster legislation, while others believe it could restrict innovation and growth.
Massachusetts Democratic Rep. Lori Trahan urged Congress to act following warnings from researchers who have resigned from AI companies, including Coxon. She called for passage of the Frontier Act, a bipartisan bill co-authored with Republican Rep. Jay Obernolte of California. If passed, the bill would introduce new safety and transparency rules for major AI companies.
As lawmakers face a narrow window to act, Sen. Bernie Sanders invited other senators to a private meeting with AI experts this week to discuss what he called the “extraordinary dangers that AI poses for humanity.” He is also backing legislation that would ban the development of uncontrollable superintelligent systems and create a new federal agency to oversee AI.
READ: The real AI revolution is happening outside tech: Venture capitalist Ajay Mago (April 8, 2026)
Following multiple recently reported incidents in which AI models from major AI firms crossed their testing environments, Democratic Rep. Ted Lieu and Republican Rep. Nathaniel Moran introduced the “AI Kill Switch Act,” a bill that would require AI developers to build and maintain technical shutdown mechanisms, giving the federal government emergency authority to slow down or completely disable rogue AI models.
“We need to get this bill across the finish line this year,” now that we’re seeing “unauthorized hacks of other companies,” said Democratic Rep. Ted Lieu.
Meanwhile, President Donald Trump has pushed back against calls for new restrictions on the current pace of AI development.
The debate over whether AI will replace humans has moved on. It now focuses on whether companies can secure and control these models and regulate them through guardrails. With the House meeting for only one more week before the midterm elections, lawmakers plan to consider the bipartisan Ratepayer Protection Act. The legislation could provide a way for giant data centers to take responsibility for the additional electricity costs caused by their facilities rather than burdening ordinary consumers.