El-Sayed Releases Policy Proposal to Get AI Development ‘Under Control’
Por Julia Conley — Common Dreams
US Senate candidate Abdul El-Sayed said Tuesday that to stop Silicon Valley “billionaires” from unilaterally deciding how artificial intelligence will be used, he proposed a regulatory framework for the technology that mirrors federal legislation pushed by top progressives in Congress.
The Michigan Democrat, who is running against Republican former Rep. Mike Rogers, unveiled his “AI Under Control” proposal following recent warnings from AI executives regarding the technology and its capacity for “recursive self-improvement”—the process through which AI models develop more sophisticated versions of themselves.
Executives including Dario Amodei of Anthropic and Sam Altman of OpenAI have said in recent weeks that AI development should be “paced” to safeguard the future of humanity, and as Common Dreams reported last week, some of those same CEOs are already plotting how to respond to a public revolt in the event of a “large-scale event” perpetrated by their tech, such as a cyberattack.
Echoing calls by lawmakers like Sen. Bernie Sanders (I-Vt.) and Rep. Greg Casar (D-Texas), El-Sayed shared a different vision, titled “AI Under Control.”
The policy proposal calls for a ban on recursive self-improvement and a new Cabinet-level government agency that would inspect AI labs, “block releases” of dangerous AI models like those that have hacked a number of websites and platforms during testing, and “order a pause or command a shutdown” when such attacks take place.
“The answer to what is stopping rogue AI agents from triggering a global nuclear crisis or deploying bioweapons cannot be the whims of a few billionaires in California,” El-Sayed said Tuesday.
The Senate candidate, a former Detroit public health official, also called for mandatory reporting on safety breaches like the one that involved OpenAI’s agents at the AI platform Hugging Face earlier this year, and for whistleblower protections.
OpenAI has been slow to publicly disclose its breaches and to notify those affected, as in the case of the company’s agents hacking an Australian government healthcare website earlier this year. The company also reportedly fired several safety researchers this month.
“The answer to what is stopping rogue AI agents from triggering a global nuclear crisis or deploying bioweapons cannot be the whims of a few billionaires in California.”
El-Sayed called for a national plan to respond to an emergency caused by AI, and reiterated his call from last month to pause advanced AI development until a government agency is established to regulate the technology, and to halt the development of AI data centers, which some communities in Michigan have fought back against.
President Donald Trump and other Republican leaders have called for rapid development of AI and a nationwide buildout of AI, fixated on the idea that, as House Speaker Mike Johnson (R-La.) said last month, “China will overlap us” if the US imposes regulations at the federal and state level.
Rogers has broken with Trump somewhat, backing a one-year moratorium on AI data centers in Michigan, but he has been criticized for holding at least $1.7 million in investments in various companies that stand to benefit from AI and data center development.
The former congressman has also accepted nearly $150,000 in campaign donations from people with AI and data center investments, including Cameron and Tyler Winklevoss and several executives from the private equity firm Blackstone.
“I understand that there’s a lot of money being sloshed around by AI super PACs, but this is the existential crisis of our time,” El-Sayed told The New York Times Tuesday. “If we don’t act quickly, we may get to a point at which we cannot act because it’s too late.”
A recent poll by the Times showed that voters in Michigan had greater confidence in El-Sayed than Rogers on the issue of AI, by an 11-point margin.
Fonte: Common Dreams
