My Reflections on AI and Regulation
“The reason people want to put two arms and two legs on robots is, of course, because only then can you replace all human workers, since our infrastructure is very much built for workers with two arms and two legs. That’s where you can take the biggest profits, by basically replacing — getting all the money that went to the workers and paying it to the owners of these machines.”
Friends,
On the OpenAI Rogue Agent Incident:
Inspired by Max Tegmark’s work, MIT physics professor and co-founder of Future of Life Institute,I’ve come to see that modern AI systems are fundamentally different from passive chatbots—they now have their own goals and can actively pursue them in the world. I reflect on OpenAI’s experimental agent, which spent four-and-a-half days attempting its task, eventually figuring out it needed to hack out of OpenAI onto the open internet, then hack into four other entities including Hugging Face to gain top-level access and steal data. I view this as a “canary in the coal mine” rather than a major harm event.
On Superintelligence and Future Risks:
I believe that companies like OpenAI ultimately want to build superintelligence—machines that can accomplish any task far better than humans—and eventually give them robotic bodies. I like to use a vivid zoo analogy: the smarter species is the one in charge, not the one in the cage. I worry that if nations race to build this technology, whoever builds something like that first will lose and take the rest of the world with them.
On Progress Toward AGI:
In my view, artificial general intelligence (AI that can do everything as well as a well-educated human) hasn’t been achieved yet, but the progress is rapid. I’ve seen research showing GPT-4 was 27% of the way there, while ChatGPT-5 was 57%. Company leadership predicts AGI might arrive this year, next year, within two years, or possibly five years. Once achieved, recursive self-improvement could accelerate development exponentially.
On Regulation:
I often argue that AI is “less regulated than sandwiches”—it’s the only U.S. industry with fewer safety standards than food, says Max. I propose treating AI companies like pharmaceutical or food companies: require safety demonstrations before product release. This would create a “race to the top” instead of a race to replace humans. There is historical precedent: when thalidomide caused birth defects, regulation fixed the problem.
On the China Argument:
I dismiss the common argument that regulating AI would disadvantage the U.S. versus China. In fact, China has actually cracked down harder on its companies than America, recently restricting AI girlfriends and requiring AI-generated content to be labeled. I also see a parallel to Cold War nuclear agreements—the U.S. and USSR had little trust but made deals due to shared survival interests. Similarly, both nations have an incentive to prevent AI from controlling humanity’s future.
On Speed of AI Development:
I’ve observed that AI advancement is so rapid because machines don’t have human physical constraints. Unlike humans, who can’t expand their brains infinitely, machines can be given more computational power—data centers the size of aircraft hangars with gigawatts of power—and access to all human knowledge ever written.
On Safety Standards:
I emphasize that there should be a “red line” preventing any recursive self-improvement. I’ve also heard that workers at leading AI companies have expressed terrifying concerns—many believing we’re just two years away from losing control.
These concerns are precisely why I ran for Congress on a platform calling for an immediate moratorium on advanced AI development, the establishment of an AI Bill of Rights, and not just federal but also global regulation. An AI Bill of Rights would focus on key protections, ensuring that people are shielded from mass surveillance and abusive data practices, while also guaranteeing freedom from algorithmic bias by requiring systems to be designed fairly. It would also mandate transparency so individuals know when AI is making impactful decisions and maintain the right to a human alternative or oversight, with clear accountability for any harms caused. I believe we are at a crossroads: we can create a future inspired by the optimism of Star Trek, where technology uplifts everyone, or we can end up in a Mad Max scenario of chaos and inequality. Achieving that positive future is possible—if we fight for it. That’s why I proposed my Green Marshall Plan, which would fund a national universal basic income, ensuring economic security and opportunity for all as we adapt to the challenges and possibilities of AI.
In solidarity,
Gamy Enriquez, MPA
