When our newly-elected legislators arrive in Washington D.C. in January, they will no doubt be asked to write new rules for artificial intelligence. So far, Congress has tried unsuccessfully to pass more than 150 AI bills over the past two sessions.
Even AI company leaders have called for more federal regulation. My worry is that most of these lawmakers will be regulating a technology they probably have never actually spent much time with and don’t really understand. That isn’t a knock on them. In my experience, it’s true of almost everyone.
In classes at Stanford and Harvard, I train people every day to use AI at work. I’ve taught nuclear contractors, national park rangers and NBA front offices. At the start of every new class, I ask my students about their general feelings around AI. Of the more than 7,000 students who have taken my class, 96 percent tell me they believe AI can help them at work. But 71 percent say they don’t know how to apply AI in their jobs. Half can’t come up with a single idea for how they might use it in their work.
Do any of us really believe our lawmakers have a better understanding of AI than the rest of the population? Chances are, our legislators are just as uncertain about how AI might change the way Americans work and live. Yet we are asking them to weigh everything from job loss and rising electricity bills to the existential threat of rogue AI agents.
I’ve seen a version of this problem play out in corporate America. CEOs eager to embrace AI insist that their employees start working with it, yet they don’t fully understand the technology themselves.
Before Congress writes the next generation of AI rules, lawmakers need to understand something that even many business leaders have yet to grasp: The real power of AI comes from treating it as a partner, rather than simply another technology tool.
That’s the first lesson I teach my students. Think of AI as a smart intern or a clever colleague. Give it an assignment. Ask it questions. Challenge its answers. You bring experience, expertise and judgment. AI helps you work faster and in entirely new ways. But one thing never changes: the most important thing about any interaction with AI is you.
I know small business owners who use AI to review contracts, allowing them to ask their lawyers more focused questions to save time and money. A fashion entrepreneur I’ve worked with collaborates with AI to analyze customer preferences and visualize how garments will look on different body types, allowing her to reduce waste and invest in local artisans. An NBA team leveraged the power of AI to respond faster to the needs of season ticket holders. A mother in one of my AI learning groups uploaded a decade of medical records to an AI and asked it to brainstorm hypotheses about her daughter’s mysterious health problems. That interaction led her to a researcher who finally gave her a diagnosis.
What these examples have in common is a different way of thinking about AI. Instead of treating it like a search engine or a piece of software, these people all approached it like a smart colleague, a collaborator, a brainstorming partner. They asked questions, shared ideas, challenged its suggestions and built on its responses. Humans bring the experience, creativity and judgment that AI lacks. That’s the kind of partnership lawmakers need to understand before they begin writing the rules that will govern it.
Those best positioned to show lawmakers “the how” of AI are not from Big Tech. They are the business leaders and innovators who are already working with it effectively every day. These are the voices of small business who can tell lawmakers the potential pitfalls of AI and the guardrails that can help make it safer for all of us.
What should be discussed? We all want guardrails, including clear rules on transparency, accountability and risk. They should be written so a company of 10 people can understand and follow them without needing to hire a consultant.
Second, I’d urge lawmakers to keep open AI models available to everyone. Most of us use AI through services like ChatGPT or Claude, where the companies control the technology. Open models can be downloaded and run independently, which gives people more control over how they use the technology. That means an independent contractor can use AI without sending sensitive customer data to a tech giant, and a creator can experiment with new ideas without paying a hefty fee every time she has a long brainstorming session.
There’s another benefit to open-source AI. It will help America compete in the AI race against China. If we restrict American open-source AI, we risk pushing developers toward Chinese technology instead. While we clearly need sensible safeguards to protect the public, we can accomplish that while also keeping American AI competitive.
So here is my suggestion to every CEO, business owner and anyone else working with AI: reach out to your representatives and offer to show them what you’ve learned. You can even use AI to research your newly elected representatives and senators. Or you can ask the AI to help you draft your letter, sharing your experiences about working with AI. Ask AI to challenge your thinking and anticipate the questions a skeptical congressional staffer might raise.
And to the lawmakers who will be writing the rules: Be curious. Ask these business owners, entrepreneurs and ordinary Americans to show you how they’re working with AI and the sorts of things it’s making possible. And lawmakers can put AI to work too, asking it to help them explore competing arguments and examine the potential consequences of the rules they’re considering.
What better way to understand AI as a partner than to work with it on the very questions Congress is trying to answer?