There has been a lot of talk about AI in the US, but in the last week or so, the tone has gotten sharper and more urgent. The White House has put a hold on a draft executive order that would have overridden state-level AI rules. This shows how complicated it is to govern a technology that moves so quickly and how intense the political climate is right now. People who know about the situation say that the decision to hold back shows how torn Washington is between supporting new ideas and letting states protect their residents. It also shows how determined the administration is to change the national AI landscape, even if that means getting into constitutional issues that are likely to cause a lot of disagreement.
The draft order that came out earlier this week would have allowed federal agencies to sue states over their laws and not give states certain types of funding if they have strict rules about AI. That one chance alone showed a big change. The main goal of the order was to stop states from making their own rules about AI, which AI companies have been asking for for a long time. This would have made the country work together on AI. Many tech leaders say that the fact that laws are so different from state to state makes it hard to roll out products, slows down research, and makes people less likely to invest. They think that a single federal standard is the best way to keep the US competitive in what they see as a global race to be the leader in AI.
But it’s never easy to talk about the federal government’s power to override state laws, and the reactions to the news showed that. State lawmakers from both parties have said many times that giving up all control over AI oversight could make it harder for them to meet the needs of their constituents. A lot of them think that AI isn’t just a new technology, but also a safety and consumer protection issue. They deal with real-life cases of fraud, deepfake images that shake people’s trust in their community, and new technologies that can make harmful or abusive content. For these leaders, the work their states are doing isn’t getting in the way of progress; it’s protecting their residents from immediate dangers they face every day.
The White House’s pause was not surprising given the tension, even though the idea had made it to the draft stage. An official stressed on Wednesday that talks about possible executive orders are still just that: talks, until something is officially announced. But the fact that the draft exists shows how far the government is willing to go to push the limits of federal power in order to make national AI policy easier to understand. In the past few years, I’ve seen similar policy fights over things like healthcare access, environmental rules, and data privacy. It’s a pattern we’ve seen before: innovation moves quickly, regulation has trouble keeping up, and the federal government is torn between making one national standard and letting states try out their own rules. This kind of tension is almost unique to the United States, and AI has made it worse than ever before.
According to papers looked at earlier this week, the draft order suggested that Attorney General Pam Bondi should be able to set up an AI Litigation Task Force. This group would only work to challenge state AI laws, saying that they get in the way of interstate trade or go against the powers of federal regulators. The Department of Commerce would also be told to look into AI rules at the state level and possibly cut broadband funding in states that don’t follow federal rules. That idea is similar to what lawmakers have talked about before, especially when it comes to the $42 billion Broadband Equity, Access, and Deployment program, which is often called BEAD. In fact, the Senate had already turned down a similar attempt earlier this year, voting 99 to 1 against stopping states from getting BEAD funds if they kept their own AI rules.
The statements made by state solicitors general during that earlier debate were very interesting. They said that taking away states’ power to enforce protective rules could make communities less safe. Their arguments were based on real-life experiences, not abstract political theory. They talked about real cases of identity theft, political content that was changed, and AI-generated content with kids in it. When local leaders talk about how quickly this kind of harm spreads, it’s easy to see why they don’t want to give up their power to regulate things. It made me think about how often we get excited about new technology before we fully understand what it will do. Innovation is exciting, but it can also be disruptive. The people who deal with those disruptions on the ground often see things very differently from those who shape national strategy.
This week, the issue came up again when former President Donald Trump said he supported adding similar provisions to the National Defence Authorisation Act. The NDAA is a bill that almost always passes and is usually about military and defence issues. Adding AI-related funding conditions to such an important law could make things a lot more difficult for states. Trump’s stance shows that he wants to speed up the development of AI and lower what he sees as regulatory barriers. During the “Winning the AI Race” summit earlier this year, he signed an executive order that stressed the need for the United States to stay ahead in artificial intelligence. In many ways, the stopped executive order seems like an extension of that same way of thinking.
Tech companies like Google, OpenAI, and the venture capital firm Andreessen Horowitz have all said they want the federal government to take the lead. They say that having to deal with different state rules slows down progress and makes things unclear, which makes investors and start-ups less likely to invest. Their point of view is important because they work in both national and international markets where speed and consistency are important. They also have a natural bias towards fewer rules and regulations. The country needs to find a way to balance that goal with the need to keep people safe from real dangers.
The pause on the executive order doesn’t mean the argument is over. If anything, it looks like it will start a more serious national conversation. What happens next will depend on how the government balances the need for innovation with the need for safety, constitutional authority, and political reality. AI is becoming more and more important in everyday life, which means that every choice has effects on almost every area, from education and healthcare to elections and national security. Being careful may be more important than being quick when so much is on the line.
